J Neurosurg 63:355-362, 1985

Acute hydrocephalus after aneurysmal subarachnoid

hemorrhage
JAN VAN G1JN, M.D., ALBERT HIJDRA, M.D., EEl,co F. M. WIJDICKS, M.D.,
MARINUS VERMEULEN, M.D., AND HANS VAN CREVEL, M.D.
Departments of Neurology, University Hospitals, Rotterdam, Utrecht, and Amsterdam,
The Netherlands

~" Hydrocephalus, defined as a bicaudate index above the 95th percentile for age, was found in 34 (20%) of
174 prospectively studied patients with suharachnoid hemorrhage (SAH) who survived the first 24 hours and
who underwent computerized tomography (CT) scanning within 72 hours. The occurrence of acute hydro-
cephalus was related to the presence of intraventricular blood, and not to the extent of cisternal hemorrhage.
The level of consciousness was depressed in 30 of the 34 patients. Characteristic clinical features were present
in 19 patients, including a gradual obtundation after the initial hemorrhage in 16 patients and small nonreactive
pupils in nine patients (all with a Glasgow Coma Scale score of 7 or less). In the remaining 15 patients (44%),
the diagnosis could be made only by CT scanning. After 1 month, 20 of the 34 patients had died: six from
rebleeding (four after shunting), 11 from cerebral infarction (eight after an initial improvement), and three
from other or mixed causes. Only one of nine patients in whom a shunt was placed survived, despite rapid
improvement in all immediately after shunting. The mortality rate among patients with acute hydrocephalus
was significantly higher than in those without, with the higher incidence caused by cerebral infarction (11 of
34 versus 12 of 140 cases, respectively; p < 0.001). Death from infarction could not be attributed to the extent
of cisternal hemorrhage, the use of antifibrinolytic drugs, or failure to apply surgical drainage, but could often
be explained by the development of hyponatremia, probably accompanied by hypovolemia.

KEY WORDS subarachnoid hemorrhage 9 aneurysm 9 hydrocephalus 9

cerebral infarction hyponatremia

the following criteria: 1) s y m p t o m s and signs of SAH

V
ENTRICULAR dilatation within hours or a few
days after aneurysmal subarachnoid hemor-
rhage (SAH), with dramatic improvement of
consciousness following immediate drainage of cere-
brospinal fluid (CSF), was reported in three different
studies in the early 1970's. 6"j~'j2 Despite the advent of
computerized tomography (CT), the subject has re-
ceived little attention since then, and one large retro-
spective study all but dismissed acute hydrocephalus as
a clinical problem, j8 Only recently has acute hydro-
cephalus again been recognized as a grave prognostic
factor, although clinical details were not available in
that study. 9 In this paper, we report the incidence,
clinical features, and outcome of acute hydrocephalus
in a prospective series of patients with a ruptured
intracranial aneurysm.
Clinical M a t e r i a l and M e t h o d s
Diagnostic Criteria
From November, 1977, to December, 1982, we pro-
spectively studied 174 consecutive patients who met
were present; 2) clinical examination and CT scanning
were performed within 72 hours of the first symptoms;
and 3) the presence of an aneurysm was proven by
angiography (107 patients) or autopsy (23 patients), or
strongly suggested by the pattern o f basal hemorrhage
on CT scans (44 patients), j7 Angiography was not un-
dertaken in the last group because surgical treatment
of the aneurysm was contraindicated by age (over 65
years), by an impaired level o f consciousness, or by
ischemic cerebral deficits. We excluded patients who
died within 24 hours o f the hemorrhage. All patients
were entered into a double-blind placebo-controlled
trial of tranexamic acid as a means o f preventing
rebleeding. ~9 The level of consciousness was assessed
by means of the 14-point Glasgow C o m a Scale (GCS).16
Computerized tomography scanning was carried out
on admission and was repeated every week for 1 month;
it was also performed within 6 hours of any clinical
deterioration. A diagnosis o f rebleeding or infarction
was confirmed by comparing the CT scanning or au-

.I. Neurosttr,~,. / Vohtme 63/September, 1985 355

 J. van Gijn, et al.

CT sections. We then chose the section that included

the foramina o f M o n r o or, if the foramina were be-
tween two sections, the mean value of the bicaudate
index was used. The results of two studies of subjects
in different age groups and without neurological
disease 3'7 allowed the definition of the upper limits of
normal (95th percentile) for the bicaudate index: 0.16
at age 30 years or under, 0.18 at 50 years, 0.19 at 60
years, 0.21 at 80 years, and 0.25 at 100 years. For the
purpose of this study, the upper limits were interpolated
for each half decade, and the bicaudate index on each
initial CT scan (Day 0 to 3 following hemorrhage) was
converted into a relative size. For instance, if a 62-year-
old patient had a bicaudate index of 0.18, the relative
size was calculated by dividing this figure by the upper
limit for age 60 (0.19), which gives 0.95.

FIG. 1. Diagram showing the method for measuring the
bicaudate index (A + B). A = the width of the frontal horns
at the level of the caudate nuclei; B = the diameter of the
brain at the same level.
topsy results with the findings on the previous CT
scan. 2~
Measurement o f Ventricular Size
As a linear measurement o f ventricular size, we used
the bicaudate index: that is, the width o f the frontal
horns at the level o f the caudate nuclei, divided by the
corresponding diameter of the brain (Fig. 1). In some
patients the ventricles were dilated to such an extent
that this measurement could be made in two adjacent
Severity of Hemorrhage
The a m o u n t o f intraventricular blood on the initial
CT scan was graded separately for each ventricle on a
scale of 0 to 3, as follows: 0 = no blood; 1 = sedimen-
tation of red blood cells in the posterior part; 2 =
ventricle partly filled with blood; and 3 -- ventricle
completely filled with blood. The extent of the cisternal
hemorrhage on the initial CT scan was similarly graded
from 0 to 3 for each of the following 10 cisterns: frontal
interhemispheric fissure, quadrigeminal cistern, and the
paired suprasellar cisterns, ambient cisterns, basal Syl-
vian fissures, and lateral Sylvian fissures. Three of the
authors separately graded all the initial CT scans in this
manner. Differences in judgment were settled by "ma-
jority vote" or, in a few cases where this was not
possible, by a joint review.

FIG. 2. Graphs showing the incidence of acute hydrocephalus (hatched bars) in 174 patients related to
two predisposing factors. Left: Distribution of relative ventricular size calculated by dividing the individual
bicaudate index by the 95th percentile for age. Center: Distribution of the extent of cisternal hemorrhage
(for a description of the grading system see text). Right: Distribution of the extent of intraventricular
hemorrhage (for a description of the grading system see text).

356 J. Neurosurg. / Volume 63 / September, 1985

Acute hydrocephalus after aneurysmal SAH

Statistical Analysis acute hydrocephalus; on average, the larger the ventri-

Distributions in 2 x 2 tables were analyzed with the
chi-square test (with Yates' correction if the total was
less than 100), and with Fisher's exact probability test
if the expected n u m b e r in any cell was less than five.
Combined 2 x 2 tables were analyzed with the Mantel-
Haenzel test.
Results
Incidence o f Acute Hydrocephalus and Predisposing
Factors
In 34 (20%) o f the 174 patients, ventricular size on
the initial CT scan exceeded the age-corrected 95th
percentile (Fig. 2 left). Table 1 shows the age, sex, site
of ruptured aneurysm, and time to admission for pa-
tients with and without hydrocephalus. None of the
differences between these two groups was statistically
significant.
The extent of cisternal hemorrhage was not associ-
ated with the presence o f acute hydrocephalus (Fig. 2
center). In contrast, intraventricular hemorrhage oc-
curred more often in patients with acute hydrocephalus,
but not in all of t h e m (Fig. 2 right). A score of 2 on
the grading scale of intraventricular blood usually rep-
resented sedimentation o f red blood cells in the de-
pendent parts o f the ventricular system. Higher scores
indicated frank intraventricular hemorrhage, which was
found in 22 of the 34 patients with acute hydrocephalus
versus 25 of the 140 patients with normal ventricular size
(p < 0.001).
cles the greater was the i m p a i r m e n t , although there
were some striking exceptions (Fig. 3). Sixteen patients
had been awake and oriented for a short t i m e after the
hemorrhage, and then gradually passed into various
degrees o f unresponsiveness within 1 day (13 patients)
or 2 days (three patients). In I 1 patients consciousness
had been depressed from the onset, and in three patients
the course o f events was not k n o w n because they h a d
been alone at the time o f hemorrhage.
Small nonreactive pupils were found in nine patients
with acute hydrocephalus, four o f w h o m also showed
persistent downward deviation o f the eyes; all other
brain-stem reflexes were intact. All nine patients with
nonreactive pupils had a relative ventricular size o f
m o r e than 1.20 and were in coma; that is, they did not
open their eyes, did not obey c o m m a n d s , and did not
utter words (GCS score of 7 or less; Fig. 3). W h e n the
occurrence of pupillary abnormalities was c o m p a r e d
with a history o f deteriorating consciousness, six pa-
tients displayed both characteristics, l0 showed pro-
gressive stupor without pupillary signs, a n d three had
small nonreactive pupils without a history o f gradually
developing c o m a (the clinical course was u n k n o w n in
two o f these three patients). Three patients with acute
hydrocephalus showed papilledema on admission,
within 24 hours of the hemorrhage; two o f these had
nonreactive pupils.

Clinical Features on Admission

The level of consciousness at the time of initial CT
scanning was impaired in 30 o f the 34 patients with

TABLE 1
Clinicalfeatures in 174 patients with and without acute
hydrocephalus following subarachnoid hemorrhage (SAH)
Acute No Acute
Clinical Feature Hydrocephalus Hydrocephalus
no. of cases 34 140
age (yrs)
range 23-78 16-83
mean _+standard deviation 55.3 • 13.4 51.7 + 15.2
sex
males 12 (35%) 54 (39%)
females 22 (65%) 86 (61%)
time SAH to admission
0-24 hrs 20 (58%) 83 (59%)
25-48 hrs 7 (21%) 37 (27%)
49-72 hrs 7 (21%) 20 (14%)
site of ruptured aneurysm*
anterior communicating artery 7 (21%) 46 (33%) FIG. 3. Relationship between degree of ventricular dilata-
carotid artery 10 (29%) 39 (28%) tion and level of consciousness as determined by the Glasgow
middle cerebral artery 2 (6%) 18 (13%) Coma Scale in 34 patients with acute hydrocephalus. Trian-
posterior circulation 4 (12%) 4 (3%) gles = patients with small nonreactive pupils; circles = other
unknown 11 (32%) 33 (23%) patients; black symbols = patients who later died from cere-
* Site of rupture was proven by angiography or autopsy. bral infarction.

J. Neurosurg. / Volume 63 / September, 1985 357

 J. van Gijn, et al.

TABLE 2
Initial clinical features in 34 patients with or without early
shunt placement/or acute hydrocephalus

Clinical Feature Shunt No Shunt

no. of cases 9 25
age (yrs)
range 46-78 23-73
mean + standard deviation 61 + 11.6 53.2 __+13.6
Glasgow Coma Scale score
_< 7 5 (56%) 6 (24%)
8-10 1 (11%) 1 (4%)
FIG. 4. Computerized tomography scans in a 59-year-old
11-13 3 (33%) 14 (56%)
woman with rupture of an aneurysm of the anterior com-
14 4(16%)
municating artery. Day of hemorrhage = Day 0. Left: Scan
small nonreactive pupils 5 (56%) 4 (16%)
on Day 2, when the patient's Glasgow Coma Scale (GCS)
ventricle size on admission*
score was 12 and the relative ventricular size was 1.4. Center:
_< 1.20 2 (22%) 13 (52%)
On Day 3, after external drainage, the GCS score was 14
> 1.20 7 (78%) 12 (48%)
(normal) and ventricular size was 0.7. Right: Scan on Day
* Based on an age-related index. For a description see text. 4 showing rebleeding that led to death on Day 5.

not recorded in two. Despite this, the clinical response

Clinical Course After Ventricular Drainage
N i n e p a t i e n t s with acute h y d r o c e p h a l u s u n d e r w e n t
v e n t r i c u l a r d r a i n a g e w i t h i n 2 weeks. I n six o f these
patients, s h u n t s with a n external catheter were placed
w i t h i n 72 h o u r s o f the hemorrhage; these were replaced
a few days later by a v e n t r i c u l o a t r i a l (VA) s h u n t . The
o t h e r three p a t i e n t s had VA s h u n t s placed i n the 2nd
week. T a b l e 2 shows that s h u n t e d patients were, o n
average, i n a worse clinical c o n d i t i o n a n d h a d larger
ventricles t h a n u n o p e r a t e d patients. It s h o u l d be em-
phasized t h a t the i n d i c a t i o n s for surgical drainage
c h a n g e d d u r i n g the study period: o f the n i n e patients
w h o were in c o m a with small fixed pupils a n d a relative
v e n t r i c u l a r size o f m o r e t h a n 1.20, o n l y the last five
p a t i e n t s u n d e r w e n t s h u n t i n g w i t h i n 72 hours.
O f the n i n e surgical patients, the C S F pressure o n
i n s e r t i o n o f the v e n t r i c u l a r catheter was higher than 20
c m H 2 0 in two patients, below this level i n five, a n d
to CSF diversion was always i m m e d i a t e a n d often dra-
matic. All five p a t i e n t s who did n o t open their eyes,
obey c o m m a n d s , or speak (GCS score of 7 or less) did
so within 1 or 2 days after early ventriculostomy, w h e n
their G C S score had i m p r o v e d to 12 or 13. In the same
interval they also regained pupillary light reflexes. In
the other four surgical patients the level o f conscious-
ness was m o d e r a t e l y i m p a i r e d (GCS score of 9 to 13),
a n d became n o r m a l w i t h i n 1 day in the patient who
was operated o n early, a n d within 2, 2, a n d 10 days in
the three p a t i e n t s who had s h u n t p l a c e m e n t in the 2 n d
week. T h e size o f the v e n t r i c u l a r system had m a r k e d l y
decreased in all six patients who u n d e r w e n t a repeat
C T scan d u r i n g this period o f clinical i m p r o v e m e n t .
After 1 m o n t h , however, eight of the n i n e s h u n t e d
patients had died (Table 3), four from infarction a n d
four from rebleeding 3, 4, 8, a n d 10 days after drainage
(Fig. 4). A n e u r y s m surgery had n o t yet b e e n u n d e r t a k e n

TABLE 3
Clinical course in 1st month after SAH in 34 patients with acute hydrocephalus*

No. of Deaths
Management and Course No. of Cases Extracranial Total Deaths
Rebleeding Infarction
Causes
ventricular drainage
no. of cases 9
initial improvement then deterioration 8 (5) 4 (3) 4 (2) 0 8 (5)
sustained improvement 1 0 0 0 0
no ventricular drainaget
no. of cases 25
Group A: normal consciousnesson admission & at 1 month 4 0 0 0 0
Group B: continued impaired consciousness 4 0 0 1 1
Group C: impaired consciousness, progressivedeterioration 5 (1) 0 3 2 (1) 5 (1)
Group D: initial improvement, then deterioration 8 (2) 2 4 (1) 0 6 (1)
Group E: sustained improvement 4 (1) 0 0 0 0
total cases 34 (9) 6 (3) 11 (3) 3 (1) 20 (7)
* Numbers in parentheses indicate the number of patients with small fixed pupils, a Glasgow Coma Scale score of 7 or less, and a relative
ventricular size of more than 1.20. For a description of ventricular size index see text. SAH = subarachnoid hemorrhage.
t For a further description of groups without ventricular drainage see text.

358 J. Neurosurg. / Volume 6 3 / S e p t e m b e r , 1985

Acute hydrocephalus after aneurysmal SAH

in any of the shunted patients. Three of the four

rebleeds after shunting occurred in the first 2 weeks
after the initial hemorrhage. All three episodes of re-
bleeding occurred in patients randomized to receive
placebo treatment instead of tranexamic acid. Analysis
of the relationship between shunting and rebleeding
during the first 2 weeks in a subgroup of 15 patients
with acute hydrocephalus who were not treated with
antifibrinolytic agents or aneurysm surgery showed that
three of four shunted patients suffered a rebleed, versus
one of 11 patients without a shunt. The difference is
suggestive, but as the numbers are small it does not
reach statistical significance on two-sided testing (p =
0.06).
Clinical Course Without Ventricular Drainage
In 25 patients acute hydrocephalus was not treated
surgically, at least not during the 1st month. The clinical
course in this period could be categorized into five types
(Groups A to E, Table 3). In four patients (Group A)
the level of consciousness was normal on admission,
and remained so after 1 month. One of these four
patients had suffered a d o c u m e n t e d hemorrhage 5 years
before. Ventricular size remained constant in three
patients and increased in one, with only temporary
disorientation. In two patients, the aneurysm was suc-
cessfully clipped after 1 month.
O f the 21 patients with impaired consciousness in
whom acute hydrocephalus was treated conservatively,
none underwent aneurysm surgery. In four of these
FIG. 5. Computerized tomography scans in a 59-year-old
patients (Group B) the level of consciousness remained woman. Top Row: On Day 0 (the day of hemorrhage) the
largely unchanged (a GCS score o f 12 in three patients, patient's Glasgow Coma Scale (GCS) score was 10. Cisternal
and of 5 to 7 in the other). One patient died suddenly and intraventricular hemorrhage suggests rupture of a basilar
within the 1st m o n t h from an extracranial complica- artery aneurysm. Ventricular size is 1.5. Center Row: On
tion, and three underwent shunting after 1 month be- Day 7, after she had spontaneously improved to a GCS score
of 13, she deteriorated to a score of 5, with right hemiplegia.
cause of a further increase in ventricular size. Progres- The ventricular size was 0.8, and infarction can be seen in the
sive deterioration (Group C) occurred in five patients. territory of the left middle cerebral artery. The increased
All five died, three from cerebral infarction and two density of the blood in the posterior horns was caused by
from the combined effects of hydrocephalus and extra- sedimentation, by the decrease in ventricular volume, and
perhaps also by abnormal cerebrospinal fluid circulation.
cranial causes. Eight patients (Group D) improved Lower Row: On Day 14, the day of death, additional and
spontaneously, but then deteriorated. The cause of extensive infarction was visible in the right hemisphere, with
secondary deterioration was rebleeding in two patients ventricular shift to the left.
(both died) and cerebral infarction in six (four of whom
died, Fig. 5). Finally, spontaneous and sustained im-
provement (Group E) occurred in four patients.
A spontaneous decrease in ventricle size of more cantly exceeded the mortality rate in the patients with-
than 20% was found on serial C T scanning in six of out acute hydrocephalus (49 o f 140 cases, or 35%; p <
the 12 patients who showed clinical improvement, at 0.02). The higher mortality rate was exclusively caused
least initially (Groups D and E); in one patient, spon- by cerebral infarction: 11 patients with acute hydro-
taneous collapse of the ventricles was accompanied by cephalus had fatal infarcts (32%) versus 12 (9%) of the
the development of bilateral subdural hematomas. A 140 other patients (p < 0.00 l). Non-fatal infarcts were
decrease of ventricular size did not occur in any of the as c o m m o n in patients with acute hydrocephalus (six
nine patients without clinical i m p r o v e m e n t (Groups B of 23 cases, or 23%) as in the other patients (28 o f 128
and C), except as a result of widespread infarction and cases, or 22%).
edema. The m e a n interval from hemorrhage to death due to
infarction was 12 days for the 11 patients with acute
Acute Hydrocephalus and Cerebral Infarction hydrocephalus and l0 days for the other 12 patients.
The 20 deaths within the 1st m o n t h after SAH among C o m p u t e r i z e d t o m o g r a p h y scanning near the t i m e of
the 34 patients with acute hydrocephalus (59%) signifi- death or autopsy was performed in nine o f the 11

J. Neurosurg. / Volume 63/September, 1985 359

 J. van Gijn, et al.

patients with infarction after acute hydrocephalus, and
in all nine o f these patients the ventricular system had
become markedly compressed by ischemic brain swell-
ing (Fig. 5). The occurrence of fatal infarction did not
depend u p o n the degree o f ventricular dilatation, the
absence o f pupillary reflexes, or the level o f conscious-
ness on admission (Fig. 3).
To investigate the nature of the relationship between
acute hydrocephalus and fatal cerebral infarction, we
studied the occurrence of other factors that have been
associated with cerebral infarction after SAH. The
a m o u n t o f cisternal hemorrhage in patients with acute
hydrocephalus (Fig. 2 center) is also shown in Fig. 6,
which gives the distribution o f patients who died fol-
lowing infarction. The distribution is similar to that in
patients who survived or who died from other causes.
Antifibrinolytic treatment (tranexamic acid, 6 gm/
day during the 1st week and 4 g m / d a y for the 3 subse-
quent weeks) was given to 86 o f the 174 patients in this
study. Fatal infarction occurred in 19% o f these pa-
tients, as opposed to 8% of the patients treated with
placebo (p < 0.05; Table 4). The presence o f acute
hydrocephalus increased the incidence o f fatal infarc-
tion in both these groups (p < 0.001).
H y p o n a t r e m i a was defined as a sodium level below
135 raM/liter on at least 2 consecutive days, in the
absence of hyperlipidemia or paraproteinemia. Because
FIG. 6. Diagram showing the extent of cisternal hemor-
rhage in all patients with acute hydrocephalus, and in those
who subsequently died from infarction (black bars).
sodium levels were measured at least three times a week
in only the last 133 patients in the series, we confined
the analysis of the correlation between hyponatremia
and infarction to this group (Table 5). A significant
correlation was found between hyponatremia and fatal
infarction (p < 0.01), and also between hyponatremia
and acute hydrocephalus (p < 0.05). In contrast, there
was no significant correlation between acute hydro-
cephalus and fatal infarction in patients without hypo-
natremia.
The role o f vasospasm could not be adequately as-
sessed in this study, since angiography was performed
in only 61% of the patients, and never at the time of
recent cerebral ischemia.
Discussion
Acute hydrocephalus was c o m m o n in our series
(20%) and was almost invariably associated with im-
paired consciousness. These findings contrast with the
large retrospective study of Vassilouthis and Richard-
son,~8 who used comparable criteria. The most probable
explanation for this difference is that we studied almost
the entire spectrum of early admissions for SAH and
did not confine ourselves to a surgical series that is
inevitably biased toward patients in good clinical con-
dition. We suspect that many patients with acute hy-
drocephalus after aneurysmal SAH never reach the
neurosurgeon because they are considered to be "in a
bad grade," irrespective of the cause. Even if CT scan-
ning is performed, moderate degrees of hydrocephalus
may be overlooked, even though our study indicates
that the prognostic implications and the therapeutic
dilemmas are often similar to those in patients with
more severe degrees of ventricular dilatation.
The clinical manifestations of acute hydrocephalus
after SAH have not been systematically studied before.
Consciousness was impaired in 30 of our 34 patients,
and improved rapidly in those in whom shunts were
placed. However, lack of responsiveness after SAH can
be caused by various other complications. 4 More spe-

TABLE 4 TABLE 5
Correlation r death from cerebral infarction with acute Correlation of death from cerebral infarction with acute
hydrocephalus and anto%rinolvtic treatment hydrocephalus and hyponatremia
Hydrocephalus Total Incidence of Hydrocephalus Total
Treatment
Acute None Cases Hyponatremia Acute None Cases
tranexamic acid hyponatremia
total cases 18 68 86 total cases 14 30 44
no. deaths 7 9 16 no. deaths 7 5 12
% deaths 39 13 19 % deaths 50 17 27
placebo no hyponatremia
total cases 16 72 88 total cases 15 74 89
no. deaths 4 3 7 no. deaths 2 4 6
% deaths 25 4 8 % deaths 13 5 7
totals totals
total cases 34 140 174 total cases 29 104 133
no. deaths 11 12 23 no. deaths 9 9 18
% deaths 32 9 13 % deaths 31 9 14

360 J. Neurosurg. / Volume 63 / September, 1985

Acute hydrocephalus after aneurysmal SAH
cific for acute hydrocephalus is a history of stupor
gradually developing within 1 or 2 days, or the presence
o f small nonreactive pupils in patients with otherwise
intact brain-stem reflexes. One or both of these char-
acteristics were noted in 19 o f our patients; in the
remaining 15 cases (44%) the diagnosis could be made
only by CT scanning.
Loss of pupillary reflexes was accompanied by down-
ward deviation of the eyes in four of nine patients.
These signs (Parinaud's syndrome) are commonly as-
sociated with hydrocephalus in infants and have also
been reported in adults with subacute obstruction of
the Sylvian aqueduct. 5,~5The syndrome has been attrib-
uted by some investigators to compression of the tectal
area by dilatation and downward herniation of the
suprapineal recess, T M but we have never encountered
this phenomenon on CT scans. The most likely expla-
nation of the ocular signs is dysfunction of the periaq-
ueductal gray matter caused by expansion of the rostral
part o f the aqueduct/5 An obstructive type of hydro-
cephalus is also suggested by the association with intra-
ventricular but not with cisternal hemorrhage. On the
other hand, seven of the 34 patients with acute hydro-
cephalus showed neither pupillary abnormalities nor
detectable intraventricular blood on CT scanning, and
cisternographic studies have indicated that blockage of
CSF flow in the subarachnoid space is also common, z
The most important finding in our series was the
association between acute hydrocephalus and subse-
quent death from cerebral infarction. This association
could not be attributed to the extent of cisternal hem-
orrhage or to the use of antifibrinolytic drugs. Within
the group of patients with acute hydrocephalus, the 11
deaths from infarction were not obviously attributable
to the initial degree of ventricular dilatation, the clinical
features, or the treatment of hydrocephalus (shunt or
no shunt).
A most intriguing explanation for the role of hydro-
cephalus in producing fatal infarction is offered by the
significant correlations between acute hydrocephalus
and subsequent hyponatremia, as well as between hy-
ponatremia and death from infarction, whereas infarcts
were relatively rare in patients with acute hydrocephalus
who did not develop hyponatremia. In a previous report
we stressed the possibly deleterious effect of fluid re-
striction, which was often instituted in our patients with
hyponatremia. 2~ This regimen was based on the as-
sumption that hyponatremia is caused by inappropriate
secretion of antidiuretic hormone,~3 but a recent study
showed a reduction rather than an expansion of plasma
volume in these patients. 1~ Fluid restriction may have
further decreased a low plasma volume from natriuresis
and may have precipitated cerebral infarctionfl ~ The
results of the present study imply that ventricular dila-
tation contributes to the development of hyponatremia,
and that it is not always SAH p e r se that causes the
loss of sodium. This notion is supported by some ob-
servations in patients with hydrocephalus from other
J. Neurosurg. / Volume 63/September, 1985
causes. ~'22 It has been suggested that the key event is
enlargement o f the third ventricle, which would inter-
fere with hypothalamic function. 22
A direct reduction of cerebral blood flow by the
enlarged ventricles has been shown in animal
experiments 23 and in xenon-inhalation studies of pa-
tients with SAH. 8 Such a mechanical effect is consistent
with the initial impairment of consciousness as well as
with the improvement after spontaneous or surgical
drainage. In contrast, a second episode o f deterioration
that ends in death from infarction might well be caused
by a fall in plasma volume. In fact, hyponatremia had
developed in five of the seven patients who showed this
clinical course and in whom serial measurements o f
the sodium level were available.
It is difficult to conclude from this study whether
surgical drainage of acute hydrocephalus is beneficial.
The immediate effect of surgery was invariably impres-
sive, despite a low ventricular pressure at the time o f
operation in about one-half of the cases. Nevertheless,
of the nine patients with shunt placement, four died
from subsequent rebleeding and four from infarction.
Shunted patients were, on average, in a worse condition
than unoperated patients, and it could be expected that
the outcome would also be worse in this group. It is
more appropriate to limit the comparison to patients
with a GCS score of 7 or less, with small fixed pupils,
and with a relative ventricular size o f more than 1.20.
The n u m b e r o f such patients was small, however: all
five shunted patients died, compared with two o f four
unoperated patients (Table 3). One other study has also
concluded that shunting did not improve the outcome. 9
On the other hand, it seems illogical not to treat acute
hydrocephalus: the enlarged ventricles may contribute
to cerebral infarction indirectly by releasing a natri-
uretic factor and directly by decreasing cerebral perfu-
sion pressure. In addition, an impaired level o f con-
sciousness often leads to a delay of aneurysm surgery
and to extracranial complications such as p n e u m o n i a
and pulmonary embolism. Perhaps the risk o f post-
shunt rebleeding can be minimized by a technique o f
gradual external drainage. Infarction might be pre-
vented, both in shunted and in unoperated patients, by
maintaining an adequate plasma volume.
Acknowledgments
We wish to thank the neurosurgeons at the University
Hospital, Rotterdam, for their cooperation and for stimulating
discussions. We are grateful to Dr. K. J. van Dongen and his
collaborators for the many emergency CT scans, and to Mrs.
M. Schipper for secretarial help.
References
1. Cledes J, Clavier J, Kerbrat G: Traitement du syndrome
d'antidiurbse inapproprire par la drmrthylchlortrtracy-
cline. Sere Hop Paris 53:823-826, 1977
2. D6czi T, Nemess~nyi Z, Szegvfiry Z, et al: Disturbances
of cerebrospinal fluid circulation during the acute stage
361

of subarachnoid hemorrhage. Neurosurgery 12:435-438,
1983
3. Earnest MP, Heaton RK, Wilkinson WE, et al: Cortical
atrophy, ventricular enlargement and intellectual impair-
ment in the aged. Neurology 29:1138-1143, 1979
4. Hijdra A, van Gijn J: Early death from rupture of an
intracranial aneurysm. J Neurosurg 57:765-768, 1982
5. Koga H, Mori K, Kawano T, et al: Parinaud's syndrome
in hydrocephalus due to a basilar artery aneurysm. Surg
Neuroi 19:548-553, 1983
6. Kusske JA, Turner PT, Ojemann GA, et al: Ventriculos-
tomy for the treatment of acute hydrocephalus following
subarachnoid hemorrhage. J Nenrosurg 38:591-595,
1973
7. Meese W, Kluge W, G r u m m e T, et al: CT evaluation of
the CSF spaces of healthy persons. Neuroradiology 19:
131-136, 1980
8. Menon D, Weir B, Overton T: Ventricular size and cere-
bral blood flow following subarachnoid hemorrhage. J
Comput Assist Tomogr 5:328-333, 1981
9. Mohr G, Ferguson G, Khan M, et al: lntraventricular
hemorrhage from ruptured aneurysm. Retrospective anal-
ysis of 91 cases. J Neurosurg 58:482-487, 1983
10. Nelson PB, Self SM, Maroon JC, et al: Hyponatremia in
intracranial disease: perhaps not the syndrome o f inap-
propriate secretion of antidiuretic hormone (SIADH). J
Neurosurg 55:938-941, 1981
11. Pertuiset B, Houtteville JP, George B, et al: Dilatation
ventriculaire prrcoce et hydrocrphalie consrcutives/t la
rupture d'anrvrysmes artrriels sus-tentoriels. Neurochi-
rurgia (Stuttg) 15:113-126, 1972
12. Raimondi A J, Tortes H: Acute hydrocephalus as a com-
plication of subarachnoid hemorrhage. Surg Neurol 1:
23-26, 1973
13. Schwartz WB, Bennett W, Curelop S, et al: A syndrome
of renal sodium loss and hyponatremia probably resulting
from inappropriate secretion of antidiuretic hormone.
Am J Med 23:529-542, 1957
14. Sballat RF, Pawl RP, Jerva M J: Significance of upward
362
J. van Gijn, et al.
gaze palsy (Parinaud's syndrome) in hydrocephalus due
to shunt malfunction. J Neurosurg 38:717-721, 1973
15. Swash M: Periaqueductal dysfunction (the Sylvian aque-
duct syndrome): a sign of hydrocephalus? J Neurol Neu-
rosurg Psychiatry 37:21-26, 1974
16. Teasdale G, Jennett B: Assessment of coma and impaired
consciousness. A practical scale. Lancet 2:81-84, 1974
17. van Gijn J, van Dongen KJ: Computerized tomography
in subarachnoid hemorrhage: difference between patients
with and without an aneurysm on angiography. Neurol-
ogy 30:538-539, 1980
18. Vassilouthis J, Richardson AE: Ventricular dilatation and
communicating hydrocephalus following sponaneous
subarachnoid hemorrhage. J Neurosurg 51:341-351,
1979
19. Vermeulen M, Lindsay KW, Murray GD, et al: Antifi-
brinolytic treatment in subarachnoid hemorrhage. N Engl
J Med 311:432-437, 1984
20. Vermeulen M, van Gijn J, Hijdra A, et al: Causes of
acute deterioration in patients with a ruptured intracranial
aneurysm. A prospective study with serial CT scanning.
J Neurosurg 60:935-939, 1984
21. Wijdicks EFM, Vermeulen M, Hijdra A, et al: Hypona-
tremia and cerebral infarction in patients with ruptured
intracranial aneurysms: is fluid restriction harmful? Ann
Neuroi 17:137-140, 1985
22. Wise BL: Inappropriate secretion of ADH caused by
obstruction of ventriculoatrial shunts. J Neurosurg 28:
429-433, 1968
23. Wozniak M, McLone DG, Raimondi A J: Micro- and
macrovascular changes as the direct cause of parenchymal
destruction in congenital murine hydrocephalus. J Neu-
rosurg 43:535-545, 1975
Manuscript received November 16, 1984.
Address reprint requests to: Jan van Gijn, M.D., Depart-
ment of Neurology, University Hospital Utrecht, P.O. Box
16250, 3500 CG Utrecht, The Netherlands.
J. Neurosurg. / Volume 63 / September, 1985