2. Stroke
JULY-AUGUST 1975
VOL. 6 NO. 4
A Journal of Cerebral Circulation
Therapy Against Brain Swelling
in Stroke Patients
A RETROSPECTIVE CLINICAL STUDY O N 227 PATIENTS
BY LIVIA CANDELISE, M.D., ALVARO COLOMBO, M.D., AND
HANS SPINNLER, M.D.
Abstract: • The effectiveness of cerebral antiedema agents in stroke has been questioned. Animal and
Therapy clinical work is inconclusive about steroids and osmotic drugs. A retrospective study of a con-
Against
Brain tinuous series of 227 stroke patients treated in the acute stage (some with dexamethasone alone,
Swelling in some with dexamethasone plus hyperosmotic mannitol infusions, and some without antiedema
Stroke therapy) showed no significant difference in the ten-day survival rate. On this criterion, there is
Patients no ground for the systematic use of such agents against this type of brain swelling.
Additional Key Words cerebrovascular disease mannitol dexamethasone
D A retrospective study was undertaken to evaluate against the unselective use of steroids in acute stroke
the short-term value of antiedema therapy in stroke patients,1418 although some suggest that the treatment
patients. Brain swelling is one of the major conse- is useful.1922 With regard to osmotic drugs, urea was
quences of a stroke, coming three or four days after triedfirst,23then mannitol,2426 andfinallyglycerol.27-28
the actual infarction. It may result in cingulate and Here again, the findings are inconclusive. Now, con-
central tentorial or uncal herniations, possibly leading current treatment with dexamethasone and osmotic
during the acute stage to rostro-caudal deterioration drugs has been proposed.29
of consciousness and ultimately to death. 1 ' 2
Moreover, edema around the ischemic area may Methods
worsen the neurological deficit.3-4 Drug therapy for
PATIENTS
stroke-induced brain swelling is based on the assump-
tion that such edema is one of the crucial mechanisms The 227 patients with cerebral infarction were admitted to
of neurological worsening and death in the acute stage the neurological wards of the Clinic for Nervous and Mental
Diseases of the University of Milan (Italy) from 1965 to
of a stroke. 1974. No a priori selection of the patients, other than that
Steroids and hyperosmotic infusions are used to determined by the criteria below, has been made.
control brain swelling of varying etiology and they are The patients constitute a continuous series, encom-
claimed to do this by different mechanisms: delayed passing three different therapeutic periods (admittedly, with
intracellular fluid accumulation in the case of the some overlapping at the interposed borderlines), namely
former, and immediate extracellular fluid accumula- that of a more or less defined vasodilator drug therapy, that
tion in the case of the latter. Experimental studies on of steroids, and that of steroids plus mannitol treatment.
steroids in ischemic brain damage are inconclusive: in- During all of these periods, comparable supportive, nursing,
effective according to several studies,5"11 effective ac- physiotherapeutic, dietary and general medical care was
cording to other studies.12'13 Many clinical reports are provided. None of the patients who entered the study
received intensive care or neurosurgery.
Cerebrovascular disease samples are made up of
From the Clinic for Nervous and Mental Diseases, University of
heterogeneous subdivisions. The criteria used for inclusion
Milan, Milan, Italy. in this study were: (1) clinical evidence of a completed stroke
Reprint requests to C. L. Clinica delle Malattie Nervose e to one hemisphere, occurring up to 24 hours before admis-
Mentali dell Universita di Milano, Via Francesco Sforza, 35-20122, sion, (2) treatment starting within 24 hours following the
Milan, Italy. stroke, and (3) evidence (from history) that this was the first
Stroke, Vol. 6. July-August 1975 353
Downloaded from stroke.ahajournals.org by on November 30, 2010
3. CANDELISE, COLOMBO, SPiNNLER
major cerebrovascular event for this patient. Some patients The frequencies within each cell of the four-
were excluded because (1) the etiology was an intracranial dimension contingency table are given in table 1.
aneurysm or an arteriovenous malformation, (2) clouding of Table 2 shows the outcome of non-parametric
consciousness or death was not definitely related to cerebral analysis.
disease, and (3) the treatment used could not be classified ac- The only comparison of statistical significance
cording to our criteria. Twenty-four percent of the original
group of patients had to be excluded. was the interaction between presence and absence of
The patients treated with antiedema medication coma and survival rate. This means that the patients
received drugs according to the following schedules: who become comatose within the first 24 hours of
1. Dexamethasone. Mean dosages were 8 mg t.i.d. onset have a significantly poorer life expectancy (67%
(range, 4 to 16 mg). The drug was supplied in 250 ml of dead) than noncomatose patients (24% dead).
Ringer's solution or 5% glucose solution t.i.d. intravenously
during the first seven days. In a small number of patients the
same dosages of steroid were given intramuscularly.
2. Dexamethasone plus mannitol. The concurrent Discussion
treatment always started from the beginning of the therapy. The negative outcome of our comparisons, which
Infusions of 250 ml of 20% mannitol plus dexamethasone (as precludes any hard inferences, may be due to the
above) were administered t.i.d. for the first three or four roughness of the ten-day survival as a criterion of
days. On the following days, the patient received the same effectiveness. And yet, our sample was large enough to
therapy as before. elicit some evidence of a relationship between treat-
ment and survival, if there was one. Cerebral edema
STATISTICAL PROCEDURES
following a stroke may not be as relevant to short-
The principal items considered were: age, patients surviving term survival expectancy as is generally thought; alter-
at the tenth day following the stroke, presence or absence of
a coma (provided the coma had been noted within the first natively, it may not respond to the type of antiedema
24 hours of the cerebral event), and the pharmacological therapy we used. The first possibility is supported by
treatment. Coma was defined as impaired consciousness at Shaw's findings,1 which by no means bear out the
or below the diencephalic level.30 predominance of stroke-induced brain swelling in the
The patients were divided into three groups according death mechanism, since for only a half of his patients
to therapy: no antiedema therapy (64 patients), dex- death could be traced back to edema. Intracranial
amethasone therapy (88 patients), and dexamethasone plus pressure studies in massive hemorrhage32 likewise fail
mannitol therapy (75 patients). to find it of key importance. Further, only 21% of
Four crossed criteria were used to classify the patients Plum's33 106 hemispheric infarct patients showed a
of each group: (1) survival rate at the tenth day following the rapidly rostro-caudal deterioration of coma suggest-
stroke, (2) presence or absence of a coma, (3) age, as below ing a clearcut supratentorial growing edema, while
or above 65 years, and (4) the three treatments listed
previously. only 13% of Ng's 2 353 stroke patients had severe brain
The result was a four-dimension contingency table.31 swelling.
The second possibility corresponds with some
Results clinical evidence which suggests that edema, mainly
Of the 227 patients, 46% were women. The average due to chronic focal lesions and certainly not prevail-
age of all patients was 56.7 years. Twenty-six percent ing in ischemic brain damage, responds to steroids.34
had diabetes and 9% had chronic atrial fibrillation Otherwise, mannitol works only on the swelling that
with mitral valve disease. None of these concomitant develops in a nonischemically damaged part of the
variables had a significantly different distribution at brain.35
the chi-square analysis. Even if nearly all cerebral lesions, both acute and
TABLE 1
Distribution of 227 Stroke Patients According to a Four-Dimension Contingency Table
With coma (82) Without coma (1 45)
Therapy Age Survivor* Dead Total Survivors Dead Total
10 13 23 3 26
CO CN
None (64) <65
>65 4 6 14 5 19
Total 5 14 19 37 8 45
Dexamethasone (88) <65 5 9 14 25 5 30
>65 4 12 16 21 7 28
Total 9 21 30 46 12 58
Dexamethasone
plus mannitol (75) <65 5 5 10 15 4 19
>65 8 15 23 12 11 23
Total 13 20 33 27 15 42
354 Stroke, Vol. 6, Jvly-Auguit 1975
Downloaded from stroke.ahajournals.org by on November 30, 2010
4. THERAPY AGAINST BRAIN SWELLING IN STROKE PATIENTS
TABLE 2 vasculature obstruction produced by ischemia. J Neurosurg
30:50-54, 1969
Four-Dimension Chi-Square Analysis of the Frequencies
Given in Table 1 7. Kahn K, Pranzarone GF, Newman T: Dexamethasone treat-
ment of experimental cerebral infarction, (abstract)
Interaction! D.F. Chi-square Neurology 22:406-407, 1972
Therapy X survival X age 2 0.799 ns 8. Siegel BA, Studer RB, Potchen EJ: Effect of dexamethasone
X coma on triethyl tin induced brain edema and the early edema in
Therapy X survival X coma 2 3.605 ns cerebral ischemia. In Reulen HJ, Schurmann K (eds): Ste-
Therapy X survival X age 2 0.733 ns roids and Brain Edema. Berlin, Springer-Verlag, p 113-121,
Therapy X survival 2 1.419 ns 1972
Survival X coma 1 31.633 < 0.005 9. Donley RF, Sundt TM: The effect of dexamethasone on the
Survival X age 1 2.762 ns edema of focal cerebral ischemia. Stroke 4:148-155 (Mar-
Apr) 1973
10. Brunson B, Robertson JT, Morgan H, et al: Evaluation of
treatment methods of cerebral infarction edema. Stroke
chronic, cause some degree of swelling, it is probably 4:461-464 (May-June) 1973
an oversimplification to classify them all by the type of 11. tee MC, Mastri AR, Waltz AG, et al: Ineffectiveness of dex-
the prevailing edema and the expected phar- amethasone for treatment of experimental cerebral infarc-
macological responsiveness to the two main classes of tion. Stroke 5:216-218 (Mar-Apr) 1974
12. Bartko D, Reulen HJ, Koch H, et al: Effect of dexamethasone
antiedemic drugs.
on the early edema following occlusion of the middle
Our feeling, supported only by scattered single cerebral artery in cats. In Reulen HJ, Schurmann K (eds):
observations, is that at least for some young patients Steroids and Brain Edema. Berlin, Springer-Verlag, p 127-
(e.g., those with internal carotid occlusion or with 137, 1972
massive embolic infarction) there is often clearcut 13. Harrison MJG, Russell RWR: Effect of dexamethasone on ex-
response to mannitol plus dexamethasone treatment, perimental cerebral infarction in the gerbil. J Neurol
both in terms of level of consciousness and survival. Neurosurg Psychiat 35:520-521, 1972
Nevertheless, we think that brain swelling is not 14. Dyken M, White PT: Evaluation of cortisone in the treatment
systematically the leading factor in the acute of cerebral infarction. JAMA 162:1531-1534, 1956
prognosis of the stroke patient and therefore the an- 15. Hetzel BS, tander H, Robson HN: Immediate treatment of
apoplexy, (correspondence) Brit Med J 1:1122, 1957
tiedema therapy cannot dramatically change the mean
16. Candelise t, Spinnler H: Dexamethasone and stroke,
life expectancy in a sample of unselected stroke (correspondence) Med J Aust 5:335 (Aug) 1972
patients. 17. Tellez H, Bauer RB: Dexamethasone as treatment in
Therefore, the wise course would seem to be to cerebrovascular disease. A controlled study in intracerebral
give antiedema therapy only to selected stroke hemorrhage. Stroke 4:541-546 (July-Aug) 1973
patients, i.e., those in whom acute cerebral edema is 18. Bauer RB, Tellez H: Dexamethasone as treatment in
likely to develop. Rapid worsening of consciousness cerebrovascular disease. A controlled study in acute cerebral
and neurological deficits soon after stroke seem to be infarction. Stroke 4:547-555 (July-Aug) 1973
the most reliable clinical signs of developing edema. 19. Russek HI, Russek AS, Zohman BL: Cortisone in immediate
Continuous monitoring for midline shift and for in- therapy of apoplectic stroke. JAMA 159:102-105, 1955
20. Roberts HJ: Supportive adrenocortical steroid therapy in
tracranial hypertension and serial EMI scanning will
acute and subacute cerebrovascular accidents, with par-
possibly further help to single out the patients for an- ticular reference to brain stem involvement. J Amer Geriat
tiedema treatment. Soc 6:686-702, 1958
21. Rubenstein MK: The influence of adrenocortical steroids on
References severe cerebrovascular accidents. J Nerv Ment Dis 141:291 -
1. Shaw CM, Alvord EC, Berry RG: Swelling of the brain follow- 299, 1965
ing ischemic infarction with arterial occlusion. Arch Neurol 22. Patten BM, Mendell J, Bruun B, et al: Double-blind study of
1:161-177, 1959 the effects of dexamethasone on acute stroke. Neurology
2. Ng tKY, Nimmannitya J: Massive cerebral infarction with 22:377-383, 1972
severe brain swelling: A clinicopathological study. Stroke 23. Javid M: Urea — new use of an old agent. Reduction of in-
1:158-163 (May-June) 1970 tracranial and intraocular pressure. Surg Clin N Amer, p
3. Weinstein JD, Toy FJ, Jaffe ME, et al: The effect of dex- 907-928 (Aug) 1958
amethasone on brain tumors. Neurology 23:121-129, 1973 24. Wise Bt, Chater N: The value of hypertonic mannitol solution
4. Reulen HJ, Hadjidimos A, Schiirmann K: The effect of dex- in decreasing brain mass and lowering cerebrospinal fluid
amethasone on water and electrolyte content and on rCBF in pressure. J Neurosurg 19:1038-1043, 1962
perifocal brain edema in man. In Reulen HJ, Schurmann K 25. Shenkin HA, Goluboff B, Haft H: The use of mannitol for the
(eds): Steroids and Brain Edema. Berlin, Springer-Verlag, p reduction of intracranial pressure in intracranial surgery. J
239-252, 1972 Neurosurg 19:897-900, 1962
5. Plum F, Alvord EC Jr, Posner JB: Effect of steroids on ex- 26. Janny P, tapalus J, Chabannes J, et al: Etude comparative
perimental cerebral infarction. Arch Neurol 9:571-573, clinique et manometrique des solution hypertoniques et des
1963 corttcoides dans le traitement de I'hypertension in-
6. Cantu R, Ames A: Experimental prevention of cerebral tracranienne. Neurochir 18:121-136, 1972
Stroke, Vol. 6, July-August 1975 355
Downloaded from stroke.ahajournals.org by on November 30, 2010
5. CANDELISE, COLOMBO, SPINNLER
27. Mathew NT, Meyer JS, Rivera VM, et al: Double-blind 32. Marks JS: Apoplectic death due to intracerebral
evaluation of glycerol therapy in acute cerebral infarction. hemorrhage: An anatomic study. Neurology 10:278-280,
Lancet 2:1327-1329 (Dec) 1972 1960
28. Meyer JS, Charney JZ, Rivera VM, et ah Treatment with 33. Plum F: Brain swelling and edema in cerebral vascular dis-
glycerol of cerebral edema due to acute cerebral infarction. ease. Res Publ Assoc Res Nerv Ment Dis 41:318-348, 1966
Lancet 2:993-997 (Nov) 1971 34. Long DM, Maxwell RE, French LA: Tlie effects of
29. Hooshmand H, Quinn JC, Houff SA: Cerebrospinal fluid glucosteroids upon experimental brain edema. In Reulen HJ,
pressure changes with chemotherapy for intracerebral Schurmann K (eds): Steroids and Brain Edema. Berlin,
hemorrhage. Neurology 22:56-61, 1972 Springer-Verlag, p 65-76, 1972
30. Plum F, Posner JB: Diagnosis of Stupor and Coma. 35. Tutt HP, Pappius HM: Studies on the mechanisms of action of
Philadelphia, F. A. Davis Co, p 80, 1972 steroids in traumatized brain. In Reulen HJ, Schurmann K
31. Goodman LA: Interactions in multidimensional contingency (eds): Steroids and Brain Edema. Berlin, Springer-Verlag, p
tables. Ann Math Stat 35:632-646, 1964 147-151, 1972
CORRECTION
An error appeared in the abstract in STROKE 6:237 (Mar- baboons the effect of blood pressure and arterial Pco2 on
Apr) 1975. local cerebral blood volume was examined. The following
equation of the regression plane relating local cerebral blood
volume (LCBV), Paco2 and mean arterial blood pressure
Measurement of Local Cerebral Blood Volume in (MABP) was obtained:
Three Dimensions in Man — Kuhl DE, Reivich M, Nyary
LCBV = 2.88 + 0.049 Paco 2 - 0.013 MABP
I, Alavi A (Cerebrovascular Research Center, Hospital of
the University of Pennsylvania, Philadelphia, Pennsylvania Local cerebral blood volume was measured in a series of
19104) eight patients and values ranged from 1.80 to 4.13 ml/100
A method has been developed for the measurement of gm depending on the location within the cross-section. The
local cerebral blood volume in man with three-dimensional higher blood volumes coincided with cortical regions. In one
resolution. Transverse section imaging with an improved patient abnormal vascularization in association with a
data processing technique enables a linear relationship to be tumor was clearly identified in the LCBV scan. In another,
obtained between the counts at any point in the scan and the the reduction in LCBV caused by edema surrounding a
radioactivity in the scanned object. This makes it possible to small glioma was demonstrated in the LCBV scan which
make absolute measurements of the concentration within also showed improved regional circulation after steroid
the brain localized in three dimensions. In a series of five therapy.
356 Stroke. Vol. 6, July-August 1975
Downloaded from stroke.ahajournals.org by on November 30, 2010