J Neurosurg 76:207-211, 1992

Effect of head elevation on intracranial pressure, cerebral

perfusion pressure, and cerebral blood flow in head-
injured patients

ZEEV FELDMAN,M.D., MALCOLMJ. KANTER,M.D., CLAUDIAS. ROBERTSON,M.D.,

CHARLES F. CONTANT, PH.D., CHRISTOPHER HAYES, P.A.-C, MICHAEL A. SHEINaERG, B.S.,
CYNTHIA A. VILLAREAL,M.D., RAJ K. NARAYAN,M.D., AND ROBERT G. GROSSMAN,M.D.
Department of Neurological Surgery, Baylor Collegeof Medicine, Houston, Texas

v- The traditional practice of elevating the head in order to lower intracranial pressure (ICP) in head-injured
patients has been challenged in recent years. Some investigators argue that patients with intracranial hyperten-
sion should be placed in a horizontal position, the rationale being that this will increase the cerebral perfusion
pressure (CPP) and thereby improve cerebral blood flow (CBF). However, ICP is generally significantly higher
when the patient is in the horizontal position. This study was undertaken to clarify the issue of optimal head
position in the care of head-injured patients. The effect of 13"and 30* head elevation on ICP, CPP, CBF, mean
carotid pressure, and other cerebral and systemic physiological parameters was studied in 22 head-injured
patients. The mean carotid pressure was significantly lower when the patient's head was elevated at 30~ than
at 0* (84.3 _+ 14.5 rata Hg vs. 895 _ 14.6 ram Hg), as was the mean ICP ( 14.1 _ 6.7 rata Hg vs~ 19.7 -+ 8.3
mm Hg). There was no statistically significant change in CPP, CBF, cerebral metabolic rate of oxygen,
arteriovenous difference of lactate, or cerebrovascular resistance associated with the change in head position.
The data indicate that head elevation to 30~significantly reduced ICP in the majority of the 22 patients without
reducing CPP or CBF.

KEY WORDS "head injury 9 head elevation 9 cerebral perfusion pressure 9

cerebral blood flow 9 intracranial pressure

LEVATED intracranial pressure (ICP) or inade- dence suggesting that CPP is maximum when patients

E quate cerebral perfusion pressure (CPP) and ce-

rebral blood flow (CBF) are common causes of
secondary brain damage in head-injured patients. In
patients with raised ICP, it is a common practice to
position the patient in bed with the head elevated above
the level of the heart. Kenning, et a[., 4 reported that
elevating the head to 45* or 90* significantly reduced
ICP. However, some studies suggest that head elevation
may also lower the CPP. Durward, et al.,2 studied the
effect of head elevation at 0 ~ (horizontal), 15", 30*, or
60 ~ on ICP, systemic and pulmonary pressure, and CPP
in patients with intracranial hypertension. They con-
cluded that 15" or 30* of head elevation significantly
reduced ICP while maintaining CPP and cardiac out-
put. However, further elevation of the head to 60*
caused an increase in ICP and a significant decrease in
CPP and cardiac output. More recently, Rosner, et
a/., 16-19have emphasized the importance of adequate
CPP and have provided physiological and clinical evi-
are in the horizontal position, even though ICP is
usually higher in this position. Davenport, etal.,' as-
sessed the role of head elevation at 20", 40", or 60 ~ in
eight patients suffering from hepatic coma. They found
inconsistent effects of head elevation on ICP but a
significant decrease in mean arterial blood pressure
(MABP) and CPP at 40* and 60*. No study has yet
measured CBF when evaluating the effects of postural
changes on cerebral perfusion. As maintenance of ad-
equate CBF is the primary rationale for maintaining
adequate CPP in patients with increased ICP, we have
studied the changes in CBF, as well as changes in ICP,
CPP, and other cerebral and systemic physiological
parameters induced by changes in head position in
patients with head injury.
Clinical Material and Methods
Twenty-two head-injured patients admitted to the
neurosurgical intensive care unit at Ben Taub General

J. Neurosurg. / Volume 76 /February, 1992 907

 Z. Feldman, et al.

TABLE 1 All patients were treated according to a protocol that

Characteristics qf 22 head-injured patients* emphasized early intubation and respiratory support,
prompt evacuation of intracranial hematomas, and pre-
CharacteristicCategory No. vention of secondary insults to the brain. 13The patients
sex (M:F) 19:3 were initially ventilated to maintain a PaO2 of 100 m m
mean age (yrs) 35 Hg or above and a PaCO2 of 30 to 35 m m Hg. Intra-
age range 18-75
pathology cranial pressure greater than 20 m m Hg was treated by
subdural hematoma 8 cerebrospinal fluid (CSF) drainage, hyperventilation
cerebral contusion 5 (PaCO2 25 to 30 m m Hg), sedation, paralysis, and
intracerebralhematoma 2 administration of mannitol. Barbiturate coma was in-
epidural hematoma 2
diffuse axonalinjury 2 duced only if ICP was refractory to the above treat-
gunshot woundto head 2 ment. 3 Medications included morphine for sedation,
traumatic subarachnoidhemorrhage 1 pancuronium for paralysis, and phenytoin for seizure
initialGCS score prophylaxis. The treatment protocols were approved by
3-5 3 the Institutional Review Board of Baylor College of
6-8 14
9-12 5 Medicine, and informed consent for each patient was
CBF (ml/100 gm/min) obtained from the next of kin.
low (< 35) 7 Within 72 hours after injury, two consecutive sets of
normal (35-60) 9 physiological measurements were performed. In the
high (> 60) 6
outcome first 13 patients, the initial set of measurements was
good recovery 6 performed with the head elevated to 300; in the subse-
moderatelydisabled 7 quent nine patients, the initial set was taken at 0* head
severelydisabled 5 elevation. The patient's head elevation was changed
vegetativestate 1 after the initial set of measurements and a second set
dead 3
of measurements was taken 45 minutes later. Measure-
* GCS = GlasgowComa Scale;CBF = cerebralblood flow. ments were made of CBF (using the Kety-Schmidt
N20 techniqueS-7), ICP (measured with the transducer
Hospital between January, 1990, and January, 1991, zeroed at the level of the foramen of Monro), arterio-
were studied. The demographic characteristics of the venous oxygen difference, cerebral metabolic rate of
patients are given in Table 1. The mean age of the oxygen (CMRO2), oxygen saturation in the jugular bulb,
patients was 35 years (range 18 to 75 years). The pri- cerebrovascular resistance (CVR), PaCO2, PaO2, arte-
mary diagnosis was subdural hematoma in eight pa- riovenous difference of lactate, and MABP. The MABP
tients, cerebral contusion in five, intracerebral hema- was measured with the transducer zeroed at the level of
toma in two, epidural hematoma in two, diffuse axonal the foramen of Monro, giving an estimate of the mean
injury in two, gunshot wound to the head in two, and carotid pressure (MCP), which was then used for cal-
traumatic subarachnoid hemorrhage in one. Postresus- culation of the CPP.
citation Glasgow Coma Scale scores were 3 to 5 in three Differences were computed by subtracting the value
patients, 6 to 8 in 14, and 9 to 12 in five. At 3 months obtained when the patient was lying flat from the value
after injury, six patients bad a good recovery, seven obtained when the head was elevated to 30 ~. These
were moderately disabled, five were severely disabled, differences were evaluated using the paired t-test. Com-
one was in a permanent vegetative state, and three had parisons among subgroups of patients were made using
died (Table 1). analysis of variance of the individual differences.

TABLE 2
Effects of head elevation on cerebral and systemic pkysiologicat parameters in 22 patients*
Parameter 30~Head Elevation 0~Head Elevation MeanDifference P Value
mean carotid pressure (mm Hg) 84.3 _+14.5 89.5 -+ 14.6 -5.2 --. 8.1 0.0062
intracranialpressure(ram Hg) 14.1 + 6.7 19.7 __.8.3 -5.7 _-X5.6
- 0.0001
cerebral perfusionpressure(CPP) (mm Hg) 70.2 +__18.1 69.7 _+ 18.7 0.45 - 8.8 0.8
cerebral blood flow(CBF)(ml/100gm/min) 47.8 _+ 16.9 48.9 ---20.4 -1.03 + 10.8 0,657
cerebrovascularresistance(CPP/CBF) 1.72 _ 0.8 1.72 + 0.9 -0.0014 ---0.356 0.9844
cerebral metabolicrate of oxygen(zmol/100gm/min) 1.35 "- 0.59 1.28 + 0.53 0.075 _+0.23 0.14
arteriovenousoxygendifference(umol/ml) 1.81 _+0.63 1.72 _+0.73 0.093 + 0.35 0.23
PaCO2(mm Hg) 32.34 _+5.89 32.46 -+ 6.4 -0.1 + 1.9 0.77
arteriovenousdifferenceof lactate (umol/ml) -0.05 + 0.05 -0.05 _+0.06 -0.0002+ 0.06 0.18
oxygensaturationin jugularbulb (%) 69.9 _+9 70.3 + 10.2 -0.0043_+0.042 0.63
* Values representmean _+standarddeviation.

208 J. Neurosurg. / Volume 76 / February, 1992

Effects of head elevation in head-injured patients

FIG. 1. Graph showing cerebral and systemic physiological parameters at head elevations of 0~ and 30~ in
22 patients with head injury. At 30", the intracranial pressure (ICP) was lower in 77% of patients, higher in
14%, and the same in 9%; mean carotid pressure (MCP) was lower in 59%, higher in 23%, and the same in
18%; cerebral perfusion pressure (CPP) was lower in 55% and higher in 45%; and cerebral blood flow (CBF)
was lower in 36%, higher in 36%, and the same in 14%. See text for analysis of the significance of the
differences.

Results
The effects of head elevation on ICP, MCP, CPP,
and CBF are illustrated in Fig. 1 and are given in Table
2 for all parameters measured. The mean values (+
standard deviation) for the MCP and ICP were signifi-
cantly lower at 30 ~ head elevation than at 0": 84.3 +
14.5 m m Hg versus 89.5 + 14.6 m m Hg for MCP, and
14.1 + 6.7 mm Hg versus 19.7 + 8.3 m m Hg for ICP.
The mean values for CPP and CBF were unaffected by
head elevation: 70.2 ___ 18.1 m m Hg at 30 ~ and 69.7 +
18.7 m m Hg at 0 ~ for CPP, and 47.8 _+ 16.9 ml/100
gm/min at 30* and 48.9 _+ 20.4 ml/100 gm/min at 0 ~
for CBF. None of the other parameters showed a statis-
tically significant difference between head elevations at
30* and at 0 ~ The results were not affected by whether
the head was at 30* or at 0* for the initial measurements.
Correlation analysis revealed a significant relation-
ship between the ICP level at 0* head elevation and the
amount of change in ICP after 30* head elevation (r = FIG. 2. Correlation between the level of intracranial pres-
sure (ICP) at 0 ~ head elevation and the amount of change in
-0.5890). The higher the ICP with the patient in the ICP after head elevation to 30~ The higher the ICP level in
horizontal position (0*), the greater the reduction in the horizontal position (0~ the greater was the reduction in
ICP at head elevation of 30* (Fig. 2). The effect of head ICP at 30~ of head elevation (r = -0.5890).

J. Neurosurg. / Volume 76/February', 1992 909


TABLE
Effects of head elevation on cerebra! and systemic physiological parameters" in six patients with ICP greater than 25 mm Hg*
Parameter 30 ~ Head Elevation
mean carotid pressure (ram Hg) 74.7 _+ 14.5
intracraniaI pressure (ICP) (ram Hg) 20.3 + 3.9
cerebral perfusion pressure (CPP) ( m m Hg) 54.3 _+ 12.8
cerebral blood flow (CBF) (ml/100 g m / m i n ) 55.9 + 9.6
cerebrovascular resistance (CPP/CBF) 0.992 + 0.226
* Values represent mean + standard deviation of the mean.
elevation on ICP in the six patients who had an ICP
greater than 25 mm Hg compared with the effect ob-
served in the other 16 patients revealed that the decrease
in ICP observed was significant in the patients with
raised ICP. The mean reduction in ICP with head
elevation was - 10.33 _+ 5.9 mm Hg in these six patients,
compared with -3.9 _+ 8.6 mm Hg for the remaining
16 patients (p < 0.05) (Table 3). In these six patients,
the ICP fell from a mean of 30.7 mm Hg at 0* to 20.3
mm Hg at 30*.
Although CBF was unchanged with head elevation
to 30 ~ when the data for all 22 patients were averaged,
in five patients the CBF fell by more than 5 ml/100
gm/min with head elevation. In these five, CBF fell
from a mean of 67.2 + 26 ml/100 gm/min at 0 ~to 50.7
_+ 16.5 ml/100 gin/rain at 30 ~ (p < 0.05). Therefore,
we tried to identify physiological parameters that would
predict such a fall. The CMRO2 was also lower after
head elevation in these five patients, falling from 1.09
+ 0.5 #mol/100 gm/min at 0* to 0.9 _ 0.5 ~mol/100
gm/min at 30*. As in the entire group, MCP and ICP
were lower with the head elevated: 83.2 _ 13.2 mm Hg
versus 90.8 _ 11.9 mm Hg for MCP, and 18 +_ 3.7 mm
Hg versus 22.4 _+ 7 mm Hg for ICP. The CVR was
higher at 30* elevation than at 0": 1.51 _+ 0.56 versus
1.22 _+ 0.49, respectively. Comparison of the changes
in these five patients to those occurring in the other 17
patients revealed that the changes in CBF, CMRO2,
and CVR were significant (p = 0.0001 for CBF, p =
0.01 for CMRO2, and p = 0.0221 for CVR). Although
the decrease in CPP in these five patients was not
significantly different from that observed in the other
patients (p = 0.119), the combination of increased CVR
and decreased CPP resulted in a decrease in CBF.
Further analysis revealed that in four of these patients,
the CBF, although lower at 30 ~ head elevation, was still
within the normal range (> 35 ml/100 gm/min). In
only one of the five patients did the CBF drop below
35 ml/100 gin/rain (from 37 to 25 ml/100 gin/rain). In
this patient, the CPP was low: 59 mm Hg at 0 ~and 50
mm Hg at 30 ~ Interestingly, the ICP rose marginally
in this patient from 12 mm Hg at 0 ~ head elevation to
14 mm Hg at 30 ~ Other than the suggestion that a low
CPP may be associated with a fall in CBF with head
elevation, the five patients in whom head elevation
produced a decrease in CBF of more than 5 ml/100
gm/min had no distinguishing features.
910
Z. Feldman, et al.
3
0 ~ Head Elevation Mean Difference P Value
83.7 + 15.1 -9 _+ 5.3 0.085
30.7 _+ 3.3 -10.33 + 5.9 0.0079
53 + 17.1 1.3 + 4.8 0.5301
59.8 _+ 26.4 -3.9 -+ 17.26 0.6042
0.976 _+ 0.398 0.016 _+ 0.219 0.8639
Discussion
The ideal head position for patients with head injury
has been disputed in recent years. Rosner 16 provided
evidence that plateau waves are triggered by a fall in
CPP and that maintaining CPP above 70 to 80 mm Hg
prevented plateau waves. In patients with intracranial
hypertension, Rosner and c o w o r k e r s ls'19 advocated
keeping the patient's head at 0 ~ elevation and using
fluid expansion or pharmacological means to maintain
CPP. Our data indicate that, although the MCP is lower
at 30 ~ head elevation than at 0", CPP and CBF are not
significantly altered in most patients by head elevation
to 30*.
Because the MCP could not be obtained by direct
measurement of the internal carotid artery pressure in
the patients we studied, we used the MABP recorded
with the transducer placed at the level of the foramen
of Monro as an estimate of the MCP, a technique
commonly used to compute CPP? 8 However, the re-
sults of a study by Woischneck, et al.,22 suggest that the
fall in MCP that we observed at 30* head elevation may
be largely an artifact of this measurement technique.
In their study addressing the accuracy of CPP measure-
ments, they found the pressure in the superficial tem-
poral artery to be almost identical to the pressure in the
internal carotid artery. When evaluating the effect of
head elevation on CPP, computed using the pressure of
the superficial temporal artery, they found the CPP to
be constant when the head was elevated from 0* to 30*.
If our measurement technique implied a fall in MCP
that did not actually occur, it would suggest that the
CPP might actually increase at 30 ~ of head elevation
and explain the lack of change in CBF seen in our
patients with head elevation.
We have also demonstrated a direct relationship be-
tween the ICP level and the change in ICP after head
elevation. Our data suggest that in head-injured patients
with normal ICP, head position does not greatly affect
ICP and that these patients can be positioned either
horizontally or with the head elevated to 30 ~. However,
when the ICP is increased, elevation of the head is very
effective in reducing the ICP. This is apparently due to:
1) hydrostatic displacement of the CSF from the cranial
cavity to the spinal subarachnoid space;4 and 2) facili-
tated venous outflow from the brain.S-10'14'ls'2LThe con-
tributions of CSF displacement and changes in venous
volume to the reduction in ICP in head-injured patients
J. Neurosurg. / Volume 76/February, 1992
Effects of head elevation in head-injured patients
after head elevation may depend upon the relative size
of each intracranial compartment at the time of head
elevation, m.~vIn patients with raised ICP, the intracra-
nial compliance decreases, t~ 12,20 and smaller changes
in intracranial volume, such as those produced by CSF
displacement, will produce greater changes in ICP.
Although the overall results of the present study show
no effect of head elevation on CBF, one must bear in
mind that our measurements are of global CBF and
that regional areas of decreased CBF may exist in the
presence of normal global blood flow measurements.
Only in one patient did the CBF fall to ischemic levels
after head elevation, and the vasodilatory cascade sug-
gested by Rosner t<~7 seemed to be demonstrable. In
this patient, the low CPP (59 m m Hg) emphasizes once
more the importance of maintaining adequate levels of
CPP. The results of the present study suggest that, in
general, head-injured patients with raised ICP should
be maintained with the head elevated to 30 ~ which will
produce a fall in ICP without a significant decrease in
CBF in the majority of these patients.
References
1. Davenport A, Will EJ, Davison AM: Effect of posture on
intracranial pressure and cerebral perfusion pressure in
patients with fulminant hepatic and renal failure after
acetaminophen self-poisoning. Crit Care Med 18:
286-289, 1990
2. Durward QJ, Amacher AL, Del Maestro RF, et at: Cere-
bral and cardiovascular responses to changes in head
elevation in patients with intracranial hypertension. J
Neurosurg 59:938-944, 1983
3. Eisenberg HM, Frankowski RF, Contant CF, et at: High-
dose barbiturate control of elevated intracranial pressure
in patients with severe head injury. J Neurosurg 69:
15-23, 1988
4. Kenning JA, Toutant SM, Saunders RL: Upright patient
positioning in the management of intracranial hyperten-
sion. Surg Neurol 15:148-152, 1981
5. Kety SS: The theory and applications of the exchange of
inert gas at the lungs and tissues. Pharmaeol Rev 3:1-41,
1951
6. Kety SS, Schmidt CF: The determination of cerebral
blood flow in man by the use of nitrous oxide in low
concentrations. Am J Physiol 143:53-66, 1945
7. Kety SS, Schmidt CF: The nitrous oxide method for the
quantitative determination of cerebral blood flow in man:
theory, procedure, and normal values. J Clin Invest 27:
476-483, 1948
8. Magnaes B: Body position and cerebrospinal fluid pres-
sure. Part l: clinical studies on the effect of rapid postural
changes. J Neurosurg 44:687-697, 1976
J. Neurosurg. / Volume 76 /February, 1992
9. Magnaes B: Movement of cerebrospinal fluid within the
craniospinal space when sitting up and lying down. Surg
Neurol 10:45-49, 1978
10. Marmarou A, Shulman K, LaMorgese J: Compartmental
analysis of compliance and outflow resistance of the
cerebrospinal fluid system. 3 Neurosurg 43:523-534, 1975
11. Maser AL, Marmarou A, Ward JD, et al: Pressure-volume
index in head injury. 3 Neurosurg 67:832-840, 1987
12. Miller JD: Significance and management of intracranial
hypertension in head injury, in Ishii S, Nagai H, Brock
M (eds): Intracranial Pressure V. Berlin: Springer-Verlag,
1983, pp 44-53
13. Narayan RK: Head injury, in Grossman RG, Hamilton
WJ (eds): Principles of Neurosurgery. New York: Raven
Press, 1991, pp 235-291
14. Paulson OB, Strandgaard S, Edvinsson L: Cerebral auto-
regulation. Cerebrovasc Brain Metab Rev 2:161-192, 1990
15. Ports DG, Deonarine V: Effect of positional changes and
jugular vein compression on the pressure gradient across
the arachnoid villi granulations of the dog. J Neurosurg
38:722-728, 1973
16. Rosner MJ: Cerebral perfusion pressure: link between
intracranial pressure and systemic circulation, in Wood
JH (ed): Cerebral Blood Flow. Physiologic and Clinical
Aspects. New York: McGraw-Hill, 1987, pp 425-448
17. Rosner M J: Vasodilatatory cascade and intracranial pres-
sure waves: theory, physiology and therapy, in Miller JD,
Teasdale GM, Rowan JO, et al (eds): Intracranial Pressure
VI. New York: Springer-Verlag, 1986, pp 137-141
18. Rosner MJ, Coley IB: Cerebral perfusion pressure, intra-
cranial pressure, and head elevation. J Neurosnrg 65:
636-641, 1986
19. Rosner MJ, Daughton S: Cerebral perfusion pressure
management in head injury. J Trauma 30:933-941, 1990
20. Shapiro K, Marmarou A, Shulman K: Characterization
of clinical CSF dynamics and neural axis compliance
using the pressure-volume index: I. The normal pressure-
volume index. Ann Neurol 7:508-514, 1980
2l. Toole JF: Effects of change of head, limb and body
position on cephalic circulation. N Engl J Med 279:
307-311, 1968
22. Woischneck D, Gaab MR, Rickels E, et al: Correct meas-
urements of cerebral perfusion pressure, in Hoff JT, Betz
AL (eds): Intracranial Pressure VII. Berlin: Springer-
Verlag, 1989, pp 850-852
Manuscript received May 19, 1991.
This study was supported in part by National Institutes of
Health Grant PO 1-NS27616.
Some portions of this paper were presented at the Eighth
International Symposium on Intracranial Pressure, in Rotter-
dam, The Netherlands, in June, 1991.
Address reprint requests to: Robert G. Grossman, M.D.,
Department of Neurosurgery, Baylor College of Medicine,
One Baylor Plaza, Houston, Texas 77030.
211