Download endocrine studies in stroke patients

UMEÅ UNIVERSITY MEDICAL DISSERTATIONS
New series No 260 — ISSN 0346-6612
From the Department of Internal Medicine,
University of Umeå, Umeå, Sweden
ENDOCRINE STUDIES IN STROKE PATIENTS
AKADEMISK AVHANDLING
som med vederbörligt tillstånd av Rektorsämbetet vid
Umeå uiversitet för avläggande av medicine doktorsexamen
kommer att offentligt försvaras i sal B, 9 tr, Tandläkarhögskolan i Umeå
fredagen den 19 januari 1990, kl 09.00
av
Tommy Olsson
Umeå 1989
ABSTRACT
ENDOCRINE STUDIES IN STROKE PATIENTS
Tommy Olsson, Department of Medicine, Umeå University,
S-901 85 Umeå, Sweden
There are a number of links between the endocrine system and the
nervous system. In this study, the impact of ischemic stroke on the
endocrine system was investigated. Elderly volunteers were studied
because data regarding the influence of advanced age on endocrine
parameters were lacking.
Only small differences in pituitary-thyroid and pituitary-adrenal
hormone axes were found between two groups of elderly patients, 60 and
80 years of age. The 80-year-old age group had a lower thyrotropin
response to thyrotropin releasing hormone (TRH) and a decline in
dopamine excretion.
Patients with acute ischemic stroke showed a pronounced
hypercortisolism studied by the dexamethasone test and urine free cortisol
measurements. In multiple regression analyses, postdexamethasone
cortisol levels were positively correlated to proximity of the lesion to the
frontal pole of the brain and disorientation. Urine cortisol levels were
predicted by limb paresis, disorientation and body temperature. High
cortisol excretion was associated with a worse functional outcome.
Norepinephrine excretion was correlated to urine cortisol levels and to
motor impairment. Patients with acute stroke had elevated free thyroxin
indices. A paradoxical growth hormone response to TRH was found in the
majority of stroke patients. In a multiple regression model disorientation
was negatively correlated to thyrotropin response after TRH and positively
correlated to prolactin response. Growth hormone response to TRH was
associated with extensive paresis.
In a cohort study diabetic and non-diabetic patients were prospectively
studied after an initial stroke. Diabetes mellitus adversely influenced
survival, the risk for a recurrent stroke and myocardial infarction.
Key words: Cerebrovascular disorders, elderly, pituitary gland, thyroid
gland, adrenal glands, hydrocortisone, dexamethasone suppression test,
cortisol excretion, diabetes mellitus, prognosis.
UMEÅ UNIVERSITY MEDICAL DISSERTATIONS
New series No 260 — ISSN 0346-6612
From the Department of Internal Medicine,
University of Umeå, Umeå, Sweden
ENDOCRINE STUDIES IN STROKE PATIENTS
Tommy Olsson
Umeå 1989
Copyright © 1989 by Tommy Olsson
ISBN 91-7174-452-5
Printed in Sweden by
the Printing Office of Umeå University
Umeå 1989
To Lillemor, Sara,
Anton and Anna
CONTENTS
Abbrevations
6
Abstract
7
Original Papers
8
Introduction
9
Aims of the study
19
Subjects
20
M ethods
23
Results
28
Discussion
38
General sum m ary and conclusions
46
Acknowledgements
48
Papers
I
69
II
91
III
109
IV
131
V
149
VI
169
5
ABBREVATIONS
ACTH = adrenocorticotropic hormone
ADL = activities of daily living
AVP = arginine vasopressin
C.i. = confidence interval
CRH = corticotropin releasing hormone
CT= com puted tom ography
DA = dopamine
E = epinephrine
FT3 I = free triiodothyronine index
FT4 I = free thyroxine index
GH = grow th hormone
GHRH = grow th hormone releasing hormone
HPA = hypothalamic-pituitary-adrenal axis
HPLC = high-performance liquid chromatography
NE = norepinephrine
RIA = radioimmunoassay
St.c. = standardized coefficient
T 3 = triiodothyronine
T 4 = thyroxine
TRH = thyrotropin releasing hormone
TSH = thyrotropin
6
ABSTRACT
ENDOCRINE STUDIES IN STROKE PATIENTS
Tommy Olsson, Department of Medicine, Umeå University,
S-901 85 Umeå, Sweden
There are a number of links between the endocrine system and the
nervous system. In this study, the impact of ischemic stroke on the
endocrine system was investigated. Elderly volunteers were studied
because data regarding the influence of advanced age on endocrine
parameters were lacking.
Only small differences in pituitary-thyroid and pituitary-adrenal
hormone axes were found between two groups of elderly patients, 60 and
80 years of age. The 80-year-old age group had a lower thyrotropin
response to thyrotropin releasing hormone (TRH) and a decline in
dopamine excretion.
Patients w ith acute ischemic stroke showed a pronounced
hypercortisolism studied by the dexamethasone test and urine free cortisol
measurements. In multiple regression analyses, postdexamethasone
cortisol levels were positively correlated to proximity of the lesion to the
frontal pole of the brain and disorientation. Urine cortisol levels were
predicted by limb paresis, disorientation and body temperature. High
cortisol excretion was associated with a worse functional outcome.
Norepinephrine excretion was correlated to urine cortisol levels and to
motor impairment. Patients with acute stroke had elevated free thyroxin
indices. A paradoxical growth hormone response to TRH was found in the
majority of stroke patients. In a multiple regression model disorientation
was negatively correlated to thyrotropin response after TRH and positively
correlated to prolactin response. Growth hormone response to TRH was
associated with extensive paresis.
In a cohort study diabetic and non-diabetic patients were prospectively
studied after an initial stroke. Diabetes mellitus adversely influenced
survival, the risk for a recurrent stroke and myocardial infarction.
Key words: Cerebrovascular disorders, elderly, pituitary gland, thyroid
gland, adrenal glands, hydrocortisone, dexamethasone suppression test,
cortisol excretion, diabetes mellitus, prognosis.
7
ORIGINAL PAPERS
I.
O lsson T, V iitanen M, H ägg E, A splund K, G rankvist K,
Eriksson S, G ustafson Y. H orm ones in "young” an d "old"
elderly: p itu ita ry -th y ro id an d p itu ita ry -a d re n a l
axes.
Gerontology 1989;35:144-52.
ü.
O lsson T, V iitanen M, H ägg E, A splund K, G rankvist K,
Eriksson S, G ustafson Y. Catecholamine excretion in old age.
Subm itted.
III.
Olsson T, Åström M, Eriksson S, Forssell Å. H ypercortisolism
revealed by the dexam ethasone suppression test in patients w ith
acute ischemic stroke. Stroke 1989; in press.
IV.
Olsson T. U rinary free cortisol excretion early after ischemic
stroke. Submitted.
V.
Olsson T, A splund K, Hägg E. Pituitary-thyroid axis, prolactin
an d grow th horm one in patien ts w ith acute stroke. In
m anuscript.
VI.
Olsson T, Viitanen M, A splund K, Eriksson S, Hägg E. Prognosis
after stroke in diabetic patients. A controlled prospective study.
Diabetologia; accepted for publication.
8
INTRODUCTION
There are a num ber of links betw een the endocrine system and the
nervous system . Every horm one can act on the central nervous
system , and the secretion of virtually every horm one is regulated
directly or indirectly by the brain (1). In addition, m any horm ones are
p roduced locally in the brain. Brain lesions m ay thus influence
endocrine function profoundly.
Stroke, or cerebrovascular disease, is a life-threatening disorder
w ith potentially devastating consequences. A m ong the survivors,
im paired life satisfaction including long-standing depressions are
com m on (2,3). Cognitive disturbances are also frequent and they
influence to a major extent the outcome of rehabilitation efforts (4).
Endocrine disturbances in psychiatric disorders such as depression
have been considered as m arkers of cen tral n e u ro tra n sm itto r
abnorm alities that m ight give clues to the underlying nature of these
diseases (1). The literature on possible interactions between stroke and
endocrine disturbances including associations w ith em otional and
cognitive changes after stroke, has been sparse so far. Furtherm ore,
the sufferers from stroke are mostly elderly people. Thus, the effects
of the disease are superim posed upon consequences of old age in the
stroke victims. Therefore, data for comparisons have to be taken from
age-m atched populations.
Hormone changes in old age
The activity of the various peripheral endocrine organs is u n d er the
influence of trophic hormones released by the pituitary. The anterior
pituitary is controlled by hormones from the hypothalam us secreted
into the portal vessels, while the hormones in the posterior pituitary
travel dow n the pituitary stalk directly from the hypothalam ic nuclei
from which they are secreted. Inform ation from suprahypothalam ic
centers are relayed by neurotransm itters to the hypothalam us. The
secretion of hypothalam ic and p itu itary horm ones is, to a major
extent, influenced by negative feed-back systems. An example of this
is the hypothalam ic-pituitary-adrenal (HPA) axis (Fig. 1), w here
cortisol exerts a negative feed-back on p itu itary and hypothalam ic
9
cells and renders them less sensitive to stim ulation. Furtherm ore,
steroid horm one receptors are found in suprahypothalam ic regions,
e.g. the hippocam pus (5).
CNS
Neurotransmittors
(5HT, Ach, NE, GABA .... )
Hypothalamus
CRH
Pituitary
ACTH
CORTISOL
Fig. 1 . The hypothalamicpituitary-adrenal (HPA) axis.
5HT=serotonin, Ach=acetylcholine, NE=norepinephrine
GABA=gamma aminobutyric
arid (modified from Til).
Adrenal
Changes w ith age at all these levels of horm one release and action
can influence the activity of various endocrine systems. An increase
of fibrosis and a decline in size of v arious organs is seen w ith
increasing age and this is also true for the endocrine system (6).
H ow ever, there are major differences w ithin the group of elderly
individuals today. Physical an d m ental w ell-being can be quite
different betw een the "young elderly" (i.e. 65-75 years old) and the
"old elderly" (75+; 7). Interference w ith non-endocrine diseases and a
reduced num ber of "old elderly in earlier studies have m ade
interpretation of data regarding age influences on the endocrine
system difficult (6). Furtherm ore, illness in old age is not seldom
10
characterized by vague and atypical sym ptom s, perhaps w rongly
attrib u ted to the aging process (8). This is particu larly true for
endocrine disorders in old age and makes it im portant to estim ate
reference intervals for horm ones in these age groups (8). H orm one
levels m ight also differ betw een sexes in old age and m ight be
influenced by confounding factors such as body mass, kidney function
and smoking habit. This has been sparsely docum ented.
It is im portant to remember that blood and urine concentrations of
h orm ones rep resen t the net resu lt of a n u m b er of d ifferen t
physiological processes. Changes w ith age in, e.g., horm one secretion
rate and distribution space, h orm one carrier pro tein s, horm one
degradation and excretion, rhytmic secretory cycles, ratios of biological
activity to im m uno-reactivity, sensitivity of horm one secretion to
feed-back regulation, horm one receptors and postreceptor events can
influence the test results (9). Much of our know ledge in these aspects,
however, is based on anim al experiments, and one m ust be cautious
w hen extrapolating findings from animal experiments to hum ans.
The hypothalamic-pituitary-adrenal axis
M orphological studies of the hypothalam us in healthy aging people
regarding the content an d tu rn o v er of h y p othalam ic releasing
horm ones or neurotransm ittors affecting horm one secretion have
given conflicting results, although the dopam ine-producing neurons
m ay be one of the most age-sensitive neurons in the central nervous
system (10,11). A num ber of morphological changes takes place w ith
age in the anterior pituitary: patchy fibrosis, focal necrosis, vascular
alterations, iron deposition and (micro)adenoma form ation are seen
(6). Furtherm ore, the gland decreases som ew hat in size (6). Despite
this, the content of the various hormones has not been reported to be
decreased (6). The reactivity of the gland regarding the HPA axis, i.e.
the adrenocorticotropin (ACTH) response to corticotropin releasing
horm one (CRH), seems to be preserved w ith increasing age (12).
There are striking effects of aging on adrenal anatom y. Cortical
nodule formation, fibrosis and pigm ent deposition are increased (13).
Biochemically, changes in hepatic steroid metabolism and a slowed
metabolic disposal probably lead to a decreased cortisol secretion rate
w ith increasing age (6). Plasma concentrations of cortisol have mostly
11
been rep o rted to be u n altered (6 ); cortisol responses to ACTH
stim ulation are also unaltered (14,15) In contrast to this, there is a
progressive decline in the secretion of d eh y d ro ep ian d ro stero n e
(DHEA) and DHEA sulfate, the major androgen metabolites from the
adrenals, and a decline in the DHEA response to ACTH and CRH in
older subjects (12,15). O lder studies have also show n an age-related
decline in 17-hydroxy- and 17-ketogenic steroid metabolites of cortisol
in urine after 40 years of age (13). Studies of u rin ary free cortisol
excretion in old age are lacking.
An increased activity of the HPA axis is a feature of the response to
various stressful stim uli (16). This stress response also includes a
heightened activity of the sym pathoadrenal system . An increased
activity of the HPA axis has been suggested as a mean of turning off
the defense reactions that are activated by stress, thereby preventing a
threat to body hom eostasis (17). The term ination of the increased
HPA activity is m ediated through the negative feed-back system as
described above.
The dexam ethasone suppression test is a screening test for
hypercortisolism used in both endocrinology and psychiatry. There
seems to be an association betw een age and post-dexam ethasone
cortisol levels in depressive patients (18). The effect of age in healthy
controls has been more debated. No association between age and the
rate of non-suppression (i. e., plasm a cortisol of > 5 p g /d l or 138
nmol/1) after dexam ethasone in healthy controls has been found. The
m ean age in these studies has in no case been over 50 years (19, for
review). In one study a positive correlation betw een old age and
plasm a cortisol after dexam ethasone, especially at 8 a. m. has been
noted (2 0 ), but in another study no such association was found (2 1 ).
Thus, the influence of age u pon the HPA axis is unclear, and
studies of healthy controls at high ages have been sparse.
The rise of plasm a norepinephrine (NE) w ith age, that has been
consistently reported (22-24), has been interpreted to be secondary to
an increase in sym pathetic tone, su p p o rted by an increase in
sym pathetic nerve firing rates in the elderly (25). No major changes
w ith age in plasm a epinephrine (E) have been rep o rted (22,24).
D ecreases as well as increases of NE and E excretion have been
reported w ith increasing age (26-28), but the num ber of study objects
over age 60 has been very lim ited and has p artly consisted of
12
institutionalized elderly.
D opam ine (DA) is a n eurotransm ittor in the central nervous
system. It is also released from peripheral nerve endings, w here it is
to a certain extent co-localized w ith NE (29). U rinary DA is probably
produced to a large extent from plasm a dopa by the action of dopa
decarboxylase in the kidney tubuli (30). The physiological role for the
p eripherally produced DA is partially unclear, b u t it has been
suggested to be an endogenous natriuretic substance (31). Studies
concerning the effect of advanced age u pon DA excretion in non­
hospitalized subjects have not been published.
Pituitary-thyroid axis including prolactin
Involutional changes are com monly seen in the thyroid gland w ith
increasing age. In an American study, only 15% of cases after age 60
w ere free from follicular and lym phoid lesions in their thyroid glands
at autopsy (32). H owever, the correlation betw een the physiological
function of the gland and various types of pathological lesions is
unclear.
In com m unity studies, a prevalence of overt hypothyroidism of
about 1-2% has been found, w ith a rising tendency w ith old age (3335). This is probably mostly due to the existence of a group of elderly
in div id u als, m ainly w om en, developing auto im m u n e th y roiditis
associated w ith increased levels of thyroid autoantibodies (34). In a
British study, overt hypothyroidism developed at a rate of 5% a year
in w om en w ho had both raised TSH concentrations an d thyroid
antibodies at the beginning of the study (36).
W hen excluding those patients w ith an overt hypothyroidism ,
there are conflicting data concerning the effect of age on serum
thyroid hormones, serum thyrotropin (TSH) and the TSH response to
th y ro tro p in releasing horm one (TRH; 34,35). The basal oxygen
consum ption rate or basal metabolic rate (BMR) decreases w ith age,
b ut BMR expressed per unit of cell mass does not change (37). The
metabolic disposal of thyroxine (T4) is reduced w ith age (38). This has
been attrib u ted to decreased p e rip h e ra l co n v ersio n of T4 to
trijodthyronine (T3). Secondary to this, the production of T3 a n d T 4
decreases w ith age (6,39). In selected
13
p o p u la tio n s
of (often
institutionalized) elderly basal levels of T 3 have repeatedly been
reported to be decreased (9).
Participation of non -h ealth y elderly in d iv id u als in p rev io u s
studies may also have obscured the interpretation of serum TSH and
its response to TRH. Some workers have found a negative correlation
betw een TSH response and age in men, b u t not in w om en (40-43),
others a low er response in elderly w om en b u t not in m en (44). Still
others have observed a low ered response in both sexes (45) or a
higher responsiveness (46). In one study of elderly m en an absence of
augm ented response to a TRH infusion test occurred despite an
increase in basal TSH; this seems to indicate an age-related decline in
the function of pituitary thyrotropes (47). Thus, conflicting data exists
regarding both age and sex influences u p o n TR H -stim ulated TSH
release in old age groups.
Basal prolactin levels have been reported to be unaltered or to
decrease around the time of m enopause in w om en (48,49). In contrast
to this, older men have been reported to have unchanged as well as
augm ented (basal and TRH-stim ulated) plasm a levels of prolactin
com pared w ith younger men (50). An equalizing tendency between
sexes in high age groups might therefore be possible.
Growth hormone (GH)
There is an age-related fall in both the num ber and the size of the
somatotropic cells in the pituitary, mostly occurring during the fourth
and fifth decades of life (51). Aging in m an is accom panied by a
generalized decrease in p ro tein sy n th esis an d a co n co m itan t
reduction in lean body mass. This suggests that grow th horm one
(GH) secretion or action may decrease w ith advancing age. In line
w ith this, a changed pulsatile pattern of basal GH release has been
reported, w ith a decline in the proportion of large GH pulses (52).
Lower levels of somatomedin-C (SM-C), the m ediator of m any of the
peripheral effects of GH, have also been reported (53,54). There are
d iv e rg e n t resu lts reg ard in g the GH an d SM-C resp o n ses to
stim ulation by grow th horm one releasing horm one (GHRH; 55-57).
However, in a study of healthy non-obese individuals, GH response
to GHRH declined w ith age w ith clearly h igher responses in
prem enopausal w om en than m en (56). GH levels in this study were
14
also positively associated w ith serum estradio l levels. This sex
difference has also been verified in 24-h profile studies of grow th
horm one secretion (53). In norm al individuals, TRH does not induce
GH secretion w hen injected intravenously (1).
Arginine Vasopressine (AVP, Antidiuretic hormone)
The vasopressin-producing cells in the hypothalam us show a slight
decrease in size w ith age until the sixth decade of life. A fter this
decade a significant increase in cell size is reported (58). Earlier studies
have given conflicting data reg ard in g b asal AVP levels w ith
increasing age (59,60).
The cortisol axis and stroke
Sufferers from stroke are potentially exposed to an acute, severe stress
situation. A fter subarachnoidal hem orrhage, an abnorm al d iu rn al
v ariatio n in plasm a cortisol an d a high incidence of abnorm al
responses to dexam ethasone have been reported (61). Furtherm ore, in
a stu d y including patien ts w ith both ischem ic an d hem orrhagic
stroke, high plasm a cortisol levels w ere associated w ith an increased
m ortality (62). The dexam ethasone test has been perform ed in a few
studies of stroke patients late after stroke. A positive association
betw een the non-suppressive state after dexam ethasone an d the
degree of handicap has been reported by som e (63,64) b u t not all
investigators (65). Cognitive disturbances as m easured by the Mini
M ental Test w ere also associated in tw o studies (63,65) w ith the
nonsuppressive state. This is in accordance w ith findings in patients
w ith chronic cognitive disturbances, such as A lzheim ers disease, in
w hich decreased suppressibility to dexam ethasone is a com m on
fin d in g (66). The d ex am eth aso n e s u p p re ssio n test has been
extensively used in psychiatry as an adjunct to clinical diagnosesetting of depression and has also been suggested as a mean to subtype
d epressions (67,68). D epressive d iso rd ers an d o th e r em o tio n al
reactions are often en co u n tered after stro k e (3,69). The n o n ­
suppressive state after dexam ethasone has been associated w ith
depression in most (63-65, 70), but not all (71) dexam ethasone studies
15
in stroke patients. The in clu d ed patien ts, how ever, have had
ischemic as well as hem orrhagic lesions at various locations and were
stu d ie d at highly various tim e-points after the acu te event.
F urtherm ore, no studies of u rin e free cortisol m easurem ents in
stroke patients are avilable. Theoretically, urine free cortisol sampling
offers several advantages over plasm a cortisol assays w hen studying
the cortisol axis in patients w ith acute stroke. Twenty-four hour urine
sam pling gives an integrative m easure of the activity in the HPA axis
(72). It probably also reflects the degree of hypercortisolism more
adequately than does plasm a cortisol m easurem ents (73). Thirdly, it
m easures the free cortisol levels, thereby avoiding possible protein
steroid-binding influences upon horm one levels (72).
Thus, there is a definite need for a m ore extensive study of the
cortisol axis in patients w ith acute stroke in a carefully selected group
of patients w here the influence of the b rain lesion in itself and
possible associations w ith depression and cognitive disturbances are
explored.
The sym pathoadrenal system and stroke
A n activation of the sym pathoadrenal system often accom panies
hypercortisolism in acute stress (74). There seem s to be direct
interactions between the HPA axis and the sym pathoadrenal system
at various levels of hormone release (74). Plasma E and NE have been
show n to be elevated in both subarachnoid hem orrhage patients and
in acute ischem ic stroke (75, for review ). E levated u rin a ry
catecholam ines have also been found in p atien ts w ith acute
cerebrovascular disease, especially in those w ith a subarachnoid
hem orrhage (76). In a recent study there was a significant positive
correlation between catecholamine excretion and blood pressure after
stroke (77). H igh catecholamine levels acting in conjunction w ith the
cortisol axis have been suggested to contribute to the ECG changes not
uncom m only seen after subarachnoid hem orrhage an d ischem ic
stroke (75,78). It w ould therefore to be of interest to stu d y the
sym pathoadrenal axis in conjunction w ith a study of the activity of
the cortisol axis.
16
Stroke and serum prolactin, grow th horm one, thyrotropin and
thyroid horm one levels
M any types of stress m ay alter an terio r horm one secretion by
influencing suprahypothalam ic control (1). Plasma levels of prolactin
and GH rise in response to hypoglycemia, exercise and surgical stress,
w hile TSH is not influenced to any major extent by these stressors
(79). Patients w ith stroke have been reported to have a decreased TSH
response to TRH (80).
A paradoxical response of grow th horm one to TRH has been
reported by Mashita et al. in a stu d y m ade at an unspecified timepoint after a cerebrovascular accident (81). In contrast, basal and
arginine-stim ulated levels of grow th horm one w ere not increased
after stroke in another study (82).
In a study group of severely ill stroke patients, Benvenga et al.
(83) found a progressive decrease in serum concentrations of T 4 and
T 3 am ong patients subsequently dying from stroke. C hanges in
thyroid horm one indices have been described in patients suffering
from a w ide variety of nonthyroidal illnesses (84,85).
Therefore, it seems to be of interest to stu d y the changes in the
pituitary-thyroid axis, prolactin an d grow th horm one secretion in
patients w ith acute stroke at a specified tim e-point after the acute
event.
Diabetes and stroke
Diabetes mellitus has been established as an independent risk factor
fo r th ro m b o em b o lic stro k e (86-90). The re la tiv e risk for
thromboembolic stroke in diabetic patients is doubled com pared with
that for non-diabetics, and stroke has been said to account for about
15% of all deaths in C aucasians w ith Type 2 diabetes (91). The
increased risk also seems to be sustained after adjustm ent for blood
p ressu re an d serum cholesterol levels (87). Preexisting cardiac
abnorm alities, such as diabetic cardiom yopathy, m ay play a role.
F u rth erm o re , changes in vessel w all m etabolism an d blood
com ponents m ay contribute to an accelerated atherosclerosis and
increased thromboembolic tendency in diabetic patients (87,92).
In an earlier study from our unit, an increased fatality rate for
17
diabetic patients during the acute phase of stroke was reported (93).
This has been verified in som e (94,95), b u t not all (96) previous
studies of short-term m ortality. W hen considering short- and long­
term m ortality after stroke, m ortality has been influenced by diabetes
in some studies (97,98), but not all (99-104).
The risk factors for a recurrent stroke after the initial event have
been sparsely studied. D iabetes seem ed to be one of the factors
associated w ith a higher m ortality due to stroke following an initial
stroke in a Russian study (94), b u t not in an A m erican com m unity
hospital study (105).
In an earlier retrospective stu d y , a shortened life span w as
dem onstrated for diabetic patients after stroke (106). The causes of
death for diabetic and non-diabetic patients dem onstrated no definite
differences but heart disorders w ere a dom inating cause of death
am ong diabetic patients in this study. D iabetes m ellitus has been
dem onstrated as a risk factor for m yocardial infarction and for a
higher m ortality in these patients, both during the acute phase and
during follow-up (107,108).
Long-term studies of a stroke patients w ith special em phasis on
the impact of diabetes mellitus as a prognostic factor are thus lacking.
Survival analysis w ith the life table technique utilizes censored
data, i.e.f individuals for w hom the event (death) has not occurred at
the time of the analysis can also be included. It is thereby possible to
study survival in a cohort of patients at variable lengths of time from
a particular starting-point, e.g. hospital adm ittance due to stroke. The
com parison betw een tw o gro u p s (diabetic versus non-diabetic
patients) can be m ade thereafter by the use of non-param etric tests
such as the logrank test. W hen determ in in g the in d ep en d en ce
influence of various risk factors u p o n survival, a m ultiv ariate
approach is appropriate. Cox's proportional-hazards model represents
a multiple regression model that is adapted for this type of data (109).
It has been suggested that hyperglycemia per se may be associated
w ith a poor outcom e after stroke (92,110). This has p artly been
attributed to the developm ent of excess lactic acidosis in the brain
after an ischemic insult in a hyperglycemic state (1 1 1 ). Hyperglycemia
has also been associated w ith the developm ent of a more pronounced
cerebral edem a in stroke patients (1 1 2 ). A nim al experim ents have,
how ever, given contradictory results, and suggestions have been
18
m ade that hyperglycem ia is not necessarily unfavorable in acute
ischemic brain dam age (113-115). Recent studies have suggested that
deleterious effects of hyperglycem ia m ay be restricted to infarcted
reg io n s h av in g c o lla te ra l c irc u la tio n , w h e re a s re g io n s of
nonanastom ising (end-arterial) vascular su p p ly m ay even benefit
from an elevated blood glucose (116).
AIMS OF THE STUDY
• To study possible changes in horm one levels betw een a "young"
and an "old" elderly age group w ith special em phasis on the
influence of sex and other confounding factors.
• To stu d y the activity of the h y p o th alam ic-p itu itary -ad ren al
axis early after stroke in a w ell-characterized group of stroke
patients, including interactions w ith the brain lesion itself and
w ith clinical sequelae to stroke such as m otor im p airm en t,
cognitive disturbances and major depression.
• To study the influence of stroke on serum thyroid hormones, and
on basal and TRH-stimulated levels of serum prolactin, GH and
thyrotropin.
• To stu d y the prognosis for diabetic p atients com pared to
non-diabetic control patients in a closed cohort prospective study
w ith a statistical technique that perm its the influence of the
diabetic disease per se to be determined.
19
SUBJECTS
Papers I and II
From the official population census register of Umeå, 167 random ly
selected m en and w om en (60 or 80 year old) that had not been
hospitalized during the last seven years were asked by postal enquiry
to participate in the study. Of these, 33 did not wish to participate and
38 did not answ er the enquiry. Therefore, the study group consisted of
96 individuals. Exclusion criteria for the studies were: central nervous
system disease including previous stroke or epilepsy, infectious
disease, thyroid disease including goiter and hyper-/hypothyroidism ,
diabetes mellitus or estrogen medication. M edication m ade exclusion
necessary from parts of the studies: cim etidin, neuroleptics and
propranolol for the pituitary-thyroid studies including prolactin;
diuretics and anti-inflammatory agents for the fluid balance study and
ß-blockers and adrenergic stim ulators for the catecholam ine study.
Exclusion due to unsatisfactory 24-h urine collection w as set to a
urin e volum e of less than 500 ml. N o in d iv idu al had a plasm a
creatinine level exceeding 150 pmol/1. After these exclusions a total of
73 persons rem ained in the first study (23 men and 16 w om en in the
60-year group, and 20 men and 14 women in the 80-year group) and 64
individuals in the second study (22 men and 13 wom en in the 60-year
group and 17 m en and 1 2 w om en in the 80-year group). Subjects
consuming at least one cigarette daily (or corresponding) were defined
as smokers.
Paper III
The patients in the study were selected from the patients adm itted to
our stroke unit. From this prospectively studied cohort, 62 patients
were selected during a 14-month period. Thirty-seven w ere m en and
25 w om en, w ith a mean±SD age of 74.6±9.4 years. Included w ere
patients w ith an acute supratentorial ischemic stroke as judged by
clinical exam ination an d CT scan. E xclusion criteria w ere:
pronounced decrease in consciousness, i.e., m ore than drow siness,
high fever (>38.5 C), renal failure (plasm a creatinine level >200
20
p.m ol/ 1), know n extensive w eight loss a n d /o r m alnutrition, hypo/h y p e rth y ro id ism , p itu ita ry insufficiency, u n co n tro lled diabetes
m ellitus, obvious abstinence reactions from alcohol a n d /o r other
central n erv o u s stim u lan ts, ep ilep sy an d certain m ed icatio n s
(g lu c o c o rtic o id s , e s tr o g e n s , a n tic o n v u ls a n ts , h ig h - d o s e
benzodiazepines, ephedrine). Twelve of the patients had know n and
well-controlled diabetes mellitus.
The first control group consisted of 25 patients (13 m en and 12
women; mean+SD age 76.8±8.4 years) acutely adm itted to the hospital
because of various acute medical disorders. N one had know n central
nervous disease including previous stroke or epilepsy. The sam e
exclusion criteria as for the stroke patients w ere applied regarding
renal function, medications, etc.
As a second control group, thirty-three random ly selected 80-yearold individuals from paper I were selected (20 m en and 13 women).
Paper IV
From the prospectively studied cohort of patients adm itted to o u r
stroke unit, 2 0 patients w ere included d u rin g a 16-month period.
There w ere 12 m en and 8 w om en w ith a mean±SD age of 77.5±8.5
years. Included were patients w ith focal neurological signs persisting
for m ore than 24 hours. They w ere considered to have an acute
supratentorial ischemic stroke as judged by clinical exam ination and
CT scan. Ten had a left-sided and 10 a right-sided b rain lesion.
Exclusion criteria were the same as in paper III. Two of the patients
had a know n diabetes mellitus. Exclusion due to unsatisfactory 24-h
urine collection w as set to a urine volum e of less than 500 ml. As a
control group, 32 of the controls from paper I w ere used (19 men and
13 women).
Paper V
The sam e inclusion criteria as for paper IV w ere used. D uring a 16m onth period 29 patients were included: 19 m en and 10 w om en w ith
a mean±SD age of 78.1±7.2 years. Thirteen had a left-sided and 16 a
right-sided brain lesion. Exclusion criteria were: pronounced decrease
in consciousness, i.e., m ore than drow siness, high fever (>38.5 C),
21
renal failure (plasma creatinine level > 2 0 0 p m o l/l), know n extensive
w eight loss a n d /o r m alnutrition, hyp o -/h y p erth y ro id ism , pituitary
insufficiency, uncontrolled diabetes m ellitus, obvious abstinence
reactions from alcohol a n d /o r other central nervous stim ulants,
epilepsy and treatm ent w ith estrogen. V arious m edications m ade
exclusion from p arts of the stu d y necessary: For p ro lactin clom ipram ine and verapam il; for grow th horm one - theophyllin.
Four patients w ith diabetes mellitus w ere excluded form the study of
grow th horm one. The same control group as in paper III was used,
i.e., thirty-three random ly selected 80-year-old individuals.
There was an overlap as to the participating subjects in papers III-V;
this is illustrated in Fig. 2.
Paper III
Paper IV
Paper V
Fig. 2.
Number of patients included in papers III-V, illustrating
the overlap of patients.
Paper VI
The patients under study (n=705) were all adm itted to our stroke unit
consecutively from January 1, 1978 to December 31, 1985. From this
population, 1 2 1 diabetic patients were identified and com pared w ith
the non-diabetic group (n=584). Each patient was followed to h is/h e r
death or until December 31, 1987. No patient was lost during follow22
up. Rates of death, recurrent stroke and m yocardial infarction w ere
recorded.
METHODS
Clinical investigations including diagnostic criteria
All participants in papers I and II w ere thoroughly investigated
clinically by a resident in internal medicine a n d /o r geriatric medicine
to exclude acute or chronic disease that could influence the test
results. The stroke patients (papers III-VI) w ere investigated in a
stan d ard ized m anner d u rin g hospital stay w ith repeated clinical
assessm ents according to the principles of our stroke u n it (117).
Inform ed consent w as obtained from all participants a n d /o r their
relatives and the studies were approved by the Ethics Com m ittee of
Umeå U niversity.
The diagnostic criteria for the various types of stroke in paper VI
w ere as follows: ( 1 ) intracerebral hem orrhage = signs of intracerebral
hem atom a on com puted tom ography (CT) scan, hem orrhagic pattern
in spinal fluid analysis, or hem orrhage in brain at autopsy, (2 ) nonembolic cerebral infarction = neurological deficits persisting m ore
than 24 hours or until death w ith no signs of bleeding on CT scan,
spinal fluid analysis or at autopsy and no potential cardiac source of
emboli, (3) embolic cerebral infarction = same as non-embolic cerebral
infarction but w ith sudden onset and a potential cardiac source of an
embolus, and (4) transient ischemic attack (TIA) = focal neurological
deficit of presum ed ischemic origin and of less than 24-h duration.
A three-point scale for orientation w as used. The agreem ent
betw een the orientation scales from tw o independent observers was
91%. The extent of paresis (in the extrem ity m ost afflicted) on the
fourth day after adm ission was quantified using a four-point scale. A
four-point mobility scale was used in papers IV and V. In a subsam ple
of patients, the results from the paresis scale and the m obility scale
correlated w ell w ith K atz' (118) ADL scores (r=0.70 an d 0.69,
respectively; pcO.OOl).
In paper VI, a four-point scale for consciousness was used. Diabetes
mellitus refers to the classification by the W orld Health O rganisation
23
(119) and h y p erte n sio n to a p rev io u sly k no w n an d trea ted
h ypertensive disease. A trial fibrillation w as diagnosed on ECG.
Congestive heart failure refers to heart insufficiency treated medically
before stroke and angina pectoris to chest discom fort elicited by
probable m yocardial ischemia. A cute m yocardial infarction w as
defined by WHO criteria (120). The clinical and autopsy criteria for
causes of death have been described by Viitanen et al. (121). Autopsy
reports w ere reviewed by two of the authors. The following registers
w ere used for the follow-up of patients: population registers w ith
dates of death, reg isters of d isch arg e diag n o ses k ep t at the
D epartm ents of Internal M edicine and G eriatric M edicine at Umeå
U niversity Hospital, hospital records, medical records kept by general
practitioners (in selected patients), and autopsy reports. A recurrent
stroke w as operationally defined as a recurrent stroke occuring more
than one w eek after the index event. Only the first recurrent stroke
was counted.
Psychiatric investigation
In paper III, the stroke patients were interviewed 4-5 days and 10 days
after adm ission by the same psychiatrist, not knowing the response to
dexam ethasone. Psychiatric diagnosis of a major depressive episode
was m ade on the basis of DSM III (Diagnostic and Statistical Manual
of Mental Disorders, ed 3), sym ptom criteria being present at both
investigations.
Laboratory m ethods
Total serum T 4 and T 3 levels w ere determ ined by com petitive
rad ioim m unoassays (RIAs) w ith an tibodies from In tern a tio n a l
Laboratory Services, London, Great Britain. The Tß-uptake test was
perform ed according to W alther (122). Free T 4 (FT 4 I) and free T3
indices (FT 3 I) w ere calculated. Serum TSH concentrations w ere
analysed by a competitive RIA and "ultrasensitive" serum TSH was
d eterm ined im m unoradiom etrically w ith a coated-tube procedure
from Farm os D iagnostica, Turku, Finland. Serum prolactin w as
analysed w ith a com petitive RIA (123) and serum grow th horm one
was analysed w ith an RIA (Phadebas hGH PRIST) from Pharm acia,
24
U ppsala, Sweden. Cortisol in plasma and urine w as analysed w ith an
RIA from Farmos Diagnostica, Turku, Finland. Plasma electrolytes
and urinary creatinine w ere m easured by stan d ard m ethods on a
m ultichannel analyzer (SMA II) from Technicon Instrum ents Corp,
Tarrytow n, NY, USA. Serum osm olality w as m easured by freezing
point depression using an OM-6010 osm om eter from Daiichi, Kyoto,
Japan. Plasma arginine vasopressin (AVP) w as analysed w ith an RIA
as described by Fyhrquist and coworkers (124). Catecholam ines in
u rine w ere concentrated by ion-pair extraction (125), separated by
HPLC on a H y p e rs il ODS 3-pm co lu m n a n d d e te c te d
electrochemically. Body mass was m easured w ith the subjects in their
underw ear and socks to the nearest 0.1 kg on a lever balance. Height
w as m easured barefoot to the nearest 1 cm. Body mass index was
calculated by dividing body w eight (kg) w ith height squared (m).
Tw enty-four-h creatinine clearance per 1.73 m 2 body surface was
calculated from the know ledge of creatinine levels in plasm a and
urine and body surface area as estimated from height and weight.
Dexam ethasone suppression test (DST)
For the elderly non-hospitalized study subjects in paper I, blood was
draw n at 8 to 8:30 a.m. for plasma cortisol analysis after an overnight
fasting. A fter taking 1 m g of dexam ethasone orally (D ecadron® ,
M erck Sharp & Dohme Internat., Rahway, NJ, USA) at 10 p. m. at
home, blood w as draw n on the following day at 8 a. m. for plasma
cortisol analysis. The stroke patien ts (papers III an d IV) w ere
investigated between the third and seventh day after adm ission to the
hospital. Blood was draw n after an over-night fast at 7 a.m for the
analysis of plasm a cortisol. The patients w ere given 1 m g of
dexam ethasone at 11 p. m. the same day. Blood w as draw n on the
following day at 7 a. m., 4 p.m. and 11 p.m. for plasm a cortisol
analyses.
Urine sam pling
In paper I a 24-hour urine collection for analysis of cortisol and
creatinine w as performed. In paper II, a 24-hour urine collection for
25
catecholam ines w as perform ed. In p ap er IV, u rin e cortisol and
catecholam ine sam pling in the stroke patients w as un d ertak en on
two consecutive days beginning betw een the first and fifth day after
adm ission to the hospital in two 24-h periods. This was done prior to
the dexam ethasone suppression test.
TRH test w ith analyses of serum TSH, prolactin and GH levels
In pap er I, a TRH-test was perform ed before lunch w ith blood
collected for TSH and prolactin analyses before, 20 min. and 60 min.
after i. v. injection of 200 pg TRH (TRH®, Hoffman-La Roche & Co.
AG, Basle, Switzerland). TSH and prolactin were analyzed as indicated
above (the com petitive TSH m ethod was used in this study). In the
stroke patients (paper V), the TRH test was perform ed betw een the
second and eigth day after adm ission to the hospital. Serum TSH,
prolactin and GH analyses were made.
Computed tomography (CT) scan of the brain
The CT scans in paper IV were analyzed by the same neuroradiologist
who was blind as to the clinical assessments. The following structural
brain m easurem ents w ere made: brain volum e was expressed as the
volum e of brain substance on three consecutive slices beginning w ith
the first slice passing through the lateral ventricles. Lesion volum e
was m easured by a com puterized calculation procedure, sum m ing the
lesion volum es in each slice w here the lesion was visible. The lesion
volum e w as divided by the overall brain volum e giving a relative
lesion volume. The distance of the lesion from the frontal pole was
determ ined by m easuring the m inim um distance of the anterio r
border of the lesion to the frontal pole. By dividing this distance by
the overall anterior-posterior length of the cerebral hem isphere in
the sam e slice, the relative lesion distance w as constructed. The
lesions w ere classified as cortical a n d /o r deep. The cortical lesions
w ere further classified w ith regard to the lobes involved (126). The
am ount of edem a surro u n d in g the lesion w as quantified using a
fo u r-p o in t scale. The an terio r horn index w as d efined as the
m axim um distance between the tips of the anterior horns divided by
the m axim um transverse inner diam eter of the skull (127). Cortical
26
brain atrophy w as estim ated using a three-point scale according to
principles previously used at our radiology d epartm ent (128). Old
lesions visible were dichotom ized into y e s/n o for each patient. In
papers IV, V and VI all patients w ere investigated by CT scan on a
routine basis as an aid to the clinical diagnosis to determ ine the type
and location of the cerebrovascular disorder.
Statistics
The statistical analyses w ere m ade in a com puterized statistical
program , SYSTAT® (129). Hormone level distributions in papers I-III
have been described by the m edian levels (m) an d 10th and 90th
percentiles. In papers I and II, w here population param eters w ere
estim ated, ninety-five percent confidence intervals (95% c.i.) were
used. They were expressed only w hen they did not include 0. Due to
scewed distributions of data and a sm all num ber of observations,
norm alization of the data w ith the use of natural logarithm s was
perform ed in papers IV and V. Pearson corrrelation coefficients were
used for the calculation of correlations. A m ultiple general linear
hypothesis program was used for one-way and m ulti-w ay analyses of
variance and covariance (papers I and II) and for the m ultiple linear
regression studies (paper III-V). The m ultiple regression studies were
perform ed w ith the use of dum m y variables (0 / 1 corresponding to
n o/yes) w hen necessery. Two-tailed t tests were utilized for testing of
the regression coefficients of each independent variable against the
dependent variable. Factor analyses w ere perform ed in papers III-V
using the principal com ponent analysis m odel in SYSTAT® w ith
varim ax rotation of the factors. These factors w ere then used in the
m ultiple regression studies. A post-hoc contrast analysis was used to
test mean differences between the stroke patients and control groups
in paper III. For the analysis of a possible prognostic inform ation of
urinary cortisol excretion (paper IV) a discrim inant analysis was used.
In paper VI, tw o-tailed t tests were used w hen app ro p riate for
differences betw een means. A m ultiple linear regression analysis
w ith Bonferroni's adjustm ent for m ultiple testing of differences in
m eans w as used to estim ate possible differences in risk factors
between the groups. Possible differences in cerebrovascular diagnoses
were calculated by a chi-square analysis. The life table technique w ith
27
Kaplan-M eier estim ate (130) w as used to assess the risk of death,
recurrent stroke and m yocardial infarction after stroke. At selected
tim e-points after stroke 95% c.i. for the life table curves w ere
estim ated. Log-rank tests were used to calculate differences between
diabetic and non-diabetic patients regarding risk of death, recurrent
stroke and myocardial infarction. M ultivariate analysis for regression
on the survival curves w as done w ith C ox's proportional hazard
m odel to determ ine the influence of the diabetic disease per se on
mortality and morbidity after stroke (109).
A p-value of <0.05 was chosen as the level of statistical significance.
RESULTS
Paper I
The results reg ard in g the influences of age an d sex u p o n the
horm one levels are sum m arized for paper I and II in Table 1.
As to the pituitary-thyroid axis, FT 4 I was unchanged w ith age, but
w om en had higher levels than m en (mean difference
1 0 .6 ,
p= 0 .0 2 ;
95% c.i. for the difference 1.8-19.4). Free T 3 index was not affected by
age or sex, and the sam e applied to conventional and ultrasensitive
serum TSH levels. The TSH response to TRH was affected by both age
and sex. Thus, the response at 20 min. was significantly higher in the
60-year-olds than in 80-year-olds (mean difference 3.2 units/1, p<0.05;
95% c.i. 0.1-6.4). This w as mostly due to the higher response for the
60-year-old w om en as com pared w ith the 80-year-old women. There
w as a significantly higher TSH response for w om en after 20 min.
(mean difference for gender was 3.9 units/1, p= 0 .0 2 ; 95% c.i. 0.7-7.1)
and after 60 min. (mean difference 3.1 units, p=0.009; 95% c. i. 0.8-5.4).
28
Table 1.
Overview of results in papers I and II. Significant differences between the age groups
and sexes are indicated, i.e., 60 = 60-year-olds had higher levels than 80-year-olds;
females = females had higher levels than males (and vice versa).
Ph
>
<i
Bä
Ph
C
HD
H
co
hJ
<C
<
DC
CO
Hi
CO
<U
c
P-.
8
CÖ
Dh
o
<
u
CO
Ö
o
N
c0o)
Dh
a>
hJ
P4
Ph
coI
CO
CO
D
C
co
HH
>H
hJ
Hi
CD
o
vo
a
V
sb
o>
W)
<
29
o
vO
a>
.5
**c
Dj
aT
ö
Dh
o
.s
O
ÏH iC
o
c C
a>M
ö
<
z
<
ou
O
co
8 •aC
aO
X
O
u
CO
a
<
a;
60
Basal prolactin levels were higher in w om en w hen the 60- and 80years- olds were taken together (mean difference 1.6 M-g/1, p=0.04). The
prolactin response to TRH was m arkedly different betw een the sexes,
women showing a two and a half times higher m edian response at 60
years and a 40% higher m edian response at 80 years after 20 min.
(mean difference betw een the sexes 19.3 |ig /l, pcO.001; 95% c.i. 11.826.8). The difference persisted after 60 min. (mean difference 13.6 pg/1,
p<0.001; 95% c.i. 8.4-18.8). No significant influence of age was found.
Males had a significantly higher urinary excretion of cortisol (mean
difference 39.9 nm ol/24h, p=0.004; 95% c.i. 12.0-66.0). There was no age
influence upon cortisol excretion.
Basal plasma cortisol levels did not differ between sexes or between
60- and 80-year-old individuals, nor did the plasm a levels after
dexam ethasone differ. The rate of non-suppression, i. e., a plasm a
cortisol >138 nmol/1 or 5(ig/dl was 4% (one 60-year-old male, one 80year-old male and one 60-year-old woman). Body mass index was
negatively correlated to the cortisol values after dexam ethasone (r= 0.26; p<0.05).
As a group, m en had slightly higher plasm a levels of AVP (mean
difference 0.65 ng/1, p=0.01; 95% c.i. 0.17-1.13).
Paper II
W hen m ales and fem ales w ere com bined, the older age group
excreted significantly less DA (m ean difference 426 n m o l/2 4 h ,
p= 0 .0 0 1 ; 95% c.i. 193-659). Eighty-year-old individuals had som ew hat
low er E and NE excretion than 60-year-old subjects. M en had a
slightly higher excretion of E and of NE and a significantly higher
excretion of DA (mean difference 252 nm ol/24h, p=0.04; 95% c.i. 14490).
Smokers excreted m ore than twice as m uch E as non-sm okers
after correction for age and sex (mean difference 26.4 nm o l/2 4 h ,
p=0.001; 95% c.i. 10.8-42.0). Smokers also had higher NE excretion
(m ean difference 93.4 n m o l/2 4 h , p=0.04; 95% c.i. 3.1-183.7).
Epinephrine excretion correlated negatively to BMI (r= -0.28, p<0.05).
A m ulti-w ay analysis of variance w ith analysis of covariance was
peform ed to determ ine the independent influence of various factors
30
u p o n h o rm o n e ex cretion. In su m m ary , DA ex cretio n w as
significantly influenced by age (standardized coefficient [s.c.]= -0.35,
p<0.05) and by creatinine clearence (st.c.=0.27, p<0.05), NE excretion
by creatinine clearence (st.c.=0.30, p<0.05), an d E excretion by
smoking (st.c.=0.31, p<0.05).
Paper III
Stroke patients had significantly higher basal cortisol levels than the
patients w ith acute m edical diseases (p= 0 .0 2 ), b u t no significant
difference w as seen between stroke patients and the healthy elderly.
M orning plasm a cortisol levels after dexam ethasone for the stroke
patients and the control groups are show n in Fig. 3. The values
w ere significantly higher in the stroke patients as com pared w ith the
healthy 80-year-olds (p=0.001). The values w ere also higher in stroke
patients as com pared w ith the patients w ith other acute disorders,
but this difference was not statistically significant (p=0.08).
In a m ultiple regression analysis right-sided lesion location and
disorientation were significantly associated w ith high cortisol levels
after dexam ethasone (st.c.=0.34 and 0.30, giving p-values 0.01 and
0.03, respectively). No influence of depression on cortisol levels was
seen.
In a second analysis of those patients w ith a visible lesion on a CT
scan, m ore anteriorly located lesions, i.e., those w ith a sh o rter
relative distance to the frontal pole, as well as low er brain volumes,
w ere associated w ith higher cortisol levels (st.c.= -0.41 and -0.32,
p=0.008 and 0.04, respectively).
Finally, the procedure w as repeated w ith the classification based
u p o n involvem ent of the v arious b rain regions. F rontal lobe
involvem ent w as associated w ith h ig h er p o st-d ex am eth aso n e
cortisol levels (st.c.=0.29, p=0.05), w hereas involvem ent of deep
structures and the occipital lobe w as associated w ith the opposite,
1.e., lower post-dexam ethasone cortisol levels (st.c.= -0.29 and -0.25,
p=0.05 and 0.09, respectively).
31
p = 0.001
8 0 0 -t
700-
60 0 -
500X
LU
Û
400-
200-
100-
*• * * :V
STROKE
C1
C2
Fig.3. Scatter plot of plasma cortisol concentration at 7 AM after
dexamethasone (DEX). STROKE, stroke patients (n=62); C l, control
patients with various acute medical disorders (n=25); C2, healthy 80-yearold controls (n=33).
32
Paper IV
U rine free cortisol levels for stroke patients and controls are shown
in Fig 4. Stroke patients had significantly higher u rin e cortisol
levels than the control group (p=0.03).
p=0.03
1000 ]
800600400-
200-
—s —
100-
806040-
20-
10 Stroke
Control
Fig. 4. Urinary free cortisol levels (logarithmic scale) in stroke patients
(n=20; mean age 77.5 years) and a control group of 80-year old volunteers
(n=32).
Plasm a cortisol levels at 7 a.m. after dexam ethasone w ere
significantly higher (p=0.003) in the stroke p atients than in the
control group. Urine free cortisol levels were significantly correlated
to cortisol levels after dexamethasone at 7 a.m. (r=0.49, p<0.001).
33
After reduction of data into orthogonal com ponents w ith a factor
analysis, a m ultiple regression analysis was m ade to determ ine the
independent influence of each com ponent u p o n urine and plasm a
cortisol levels. Urine free cortisol levels were significantly associated
w ith lim b paresis (st.c.=0.59, p=0.003), disorientation (st.c.=0.41,
p=0.03) and body tem perature (st.c.= 0.39, p=0.03). Plasma cortisol
levels also tended to be influenced by these factors, and by increased
age, but none of these factors reached statistical significance.
The data that yielded significant results in the m ultiple regression
analyses in papers III and IV are sum m arized in Table 2.
The possible prognostic value of cortisol m easurem ent w as
thereafter studied. W ith a discrim inant analysis, u rin e cortisol
values significantly predicted the functional outcom e even after
adjustm ent for age (p=0.001). Cortisol levels after dexam ethasone
were also associated w ith the functional outcome, (p= 0 .0 2 ).
U rine levels of norepinephrine w ere positively correlated to
u rine cortisol levels (r=0.54, p<0.05) and to limb paresis (r=0.52,
p<0.05), w hereas ep in ep h rin e an d d opam ine exhibited w eaker
correlations to these factors.
Table 2.
Results from the multiple regression analyses in papers No. Ill
and IV. Note: n=54 for clinical variables regarding plasma
cortisol; n-40 for CT scan variables and n=20 for urine cortisol.
Plasma kortisol after DEX
Urine cortisol
St.c.
p-value
St.c.
p-value
Disorientation
0.30
0.03
0.41
0.03
Limb paresis
0.11
0.42
0.59
0.003
Distance from
the frontal
pole to lesion
-0.41
0.008
-
-
Brain volume
-0.32
0.04
-
-
St.c. = standardized coefficient
34
Paper V
Stroke patients had slightly higher FT 4 I levels as com pared w ith
control patients (mean ± SEM 117 ± 3 vs. 106 ± 4; p = 0.01). The FT3 I
levels w ere som ew hat low er in stroke patients, as w as the TSH
response to TRH. Basal and stim ulated m ean levels of prolactin did
not differ significantly between the groups. In a m ultiple regression
model, the FT4 I level was weakly associated w ith the extent of motor
im pairm ent, w hereas the FT3 I level was not predicted by extremity
paresis. Low thyrotropin levels 20 and 60 min. after TRH w ere
associated w ith disorientation (st.c.= -0.43 an d -0.41, p=0.02 an d
p=0.04, respectively). The prolactin response to TRH w as positively
predicted by disorientation, especially after 60 min. (st.c.=0.42,
p=0.045). The extent of extremity paresis independently predicted the
GH response 60 min. after TRH (st.c.=0.50, p=0.02).
Paper VI
O ut of the 1 2 1 diabetics all but six w ere considered to have Type 2
diabetes. Diabetic patients were clearly more afflicted by heart disease
before the index event as indicated by significantly more heart failure
and angina pectoris and a trend tow ards an increased proportion of
p rev io u s m yocardial infarction. H em atocrit levels w ere also
significantly higher in diabetic (43.7±4.7%) than in nondiabetic
patients (42.1±4.7%; pcO.Ol).
W hen calculated by a chi-square analysis there was a significant
difference in cerebrovascular diagnoses (chi-square-value 11.7, df=3,
p=0.008). Diabetic patients were more commonly afflicted by cerebral
embolism and less commonly by transient ischemic attacks. Diabetic
patients also tended to show a more severe neurological im pairm ent
at onset.
In all, 416 patients died during observation: 71% (86/121) of the
diabetic and 57% (330/584) of the non-diabetic patients. M ortality
after stroke in these two groups are presented by life table curves in
Fig. 5. Ninety-five percent confidence intervals are given at selected
time-points. M edian survival time for diabetic patients was 513 days
and for non-diabetic patients 1569 days. The risk of death after stroke
35
w as increased for the diabetic group m ainly d u rin g the first six
m onths after the index event. Risk of death as calculated w ith a logrank test w as significantly higher for diabetic patients (chi-square
value, 21.0 df=l, p<0.001). Diabetic patients w ere more prone to suffer
a recurrent stroke (chi-square value 11.9, ld f, p=0.001). Diabetic
patients had a higher risk for myocardial infarction w hen calculated
by a log-rank test (chi-square value 6.7, 1 df, p=0.001), b u t the 95%
confidence intervals were overlapping.
0)
4->
«3
£
-H
-P
.0
8
U)
0)
M ,6
Q)
-H
0)
Si
c
fö
I-- 1 0 . 2
P
mh
0. 0
0
6
2
8
10
Years after stroke
Fig.5. Risk of death in diabetic (bold line) and non-diabetic (thin line)
patients analysed by the life table technique. 95% confidence intervals are
given at six months, one, two and three years after stroke.
To determ ine the independent influence from diabetes and other
risk factors u p o n m ortality, recu rren t stroke an d m yocardial
infarction, Cox's proportional hazard m odel was used. The m ain
resu lts are p resen ted in Table 3. D iabetes m ellitus as such
significantly influenced m ortality and the risk of a recurrent stroke.
Diabetes mellitus slightly increased the risk of myocardial infarction
b u t this effect did not reach statistical significance. The strongest
impact upon survival was from advanced age, followed by a lowered
level of consciousness at arrival to the hospital. H eart failure and
hypertension also influenced survival significantly. For recurrent
36
stroke, the significant risk factors besides diabetes mellitus were heart
failure an d advanced age. The co rresp o n d in g risk factors for
m yocardial infarction after stroke were angina pectoris, advanced age
and heart failure.
Table 3. Results from a multivariate analysis with Cox's proportional
hazard model for factors influencing mortality, recurrent stroke and
myocardial infarction after the initial stroke. * = p<0.05, ** = p<0.01, *** =
p<0.01. TIA=transient ischemic attack. MI=myocardial infarction.
Risk factor
t-value
Recurrent stroke
M ortality
Advanced age
8.28***
2.71**
MI
2.97**
Lowered conscious­
ness at adm ission
7.33***
H eart failure
4.32***
3.56***
2.07*
Diabetes mellitus
2.50*
2.63*
1.56
H ypertension
2.40*
1.36
0.25
A ngina pectoris
1.69
0.39
3.02**
-0.92
0 .1 0
Early after adm ission, the acute brain disease in itself causing
cerebral edem a and coning was the major cause of d eath in both
groups. D uring the period from 1 w eek to 1 m onth after stroke,
pulm onary em bolism w as the major cause of d eath in diabetic
patients, causing m ore than half of the deaths i.e., 5 out of 9 deaths.
The p ro portion of cardiac deaths after the first six m onths of
observation w as 29 and 33% for diabetics and non-diabetic patients,
respectively. The main reason for the excess m ortality in the diabetic
patients during the first 6 m onths apppears to be the cerebrovascular
lesion as such. W hen calculated by a chi-square-analysis, the
difference in the causes of death between diabetic and non-diabetic
patients did not reach statistical significance.
37
DISCUSSION
Paper I
This paper sum m arizes the effects of age, sex and of confounding
factors on certain endocrine variables in a study group of carefully
examined elderly volunteers. The utilization of this elderly stu d y
group makes it possible to deduce changes in the endocrine systems
u n d e r stu d y b etw een the ”young" an d "o ld ” eld erly state.
Furtherm ore, the use of one-way and multi-way analysis of variance
w ith the concom itant analysis of covariance m akes it possible to
draw conclusions about the independent effects of age, sex, etc. as
such in these old-age groups.
O ur results agree w ith a preserved activity in the HPA axis in old
age. No influence of age could be docum ented for cortisol excretion
and this w as also true for plasm a cortisol levels before and after
dexamethasone. The sex difference that has earlier been docum ented
in young individuals (131) regarding cortisol excretion also seems to
persist in advanced age. The negative correlation betw een BMI and
post-dexam ethasone cortisol levels points tow ard a higher level
activity in the adrenocortical system in the lean elderly. Thus, the
cortisol axis has a preserved activity in old age, supported by findings
in other studies of an unaltered response of ACTH to CRH (12) and
of cortisol to ACTH (14,15).
An age-related decline in the TSH response to TRH was found. A
reduction of thyreotrophic function w ith increasing age is supported
by the absence of an augm ented response to TRH infusion despite an
increase in basal TSH in a study m ade by H arm an et al. (47). The
physiologic significance of these results may be uncertain, as TRH is
given in large doses over a lim ited time period. The results are,
however, relevant to the diagnostic use of TRH.
In our carefully screened p o p u latio n w ith o u t overt thyroid
disease w e could not discover any influence of advanced age upon
peripheral horm one levels. It is well in line w ith other studies
utilizing healthy populations in which either no change or a modest
decrease in T 3 and T 4 levels has been found (9). A preserved reserve
capacity of the thyroid gland has also been dem onstrated during
acute stress situations in the elderly (132).
38
Prolactin horm one levels are substantially influenced by a central
dopam inergic inhibitory pathw ay. As dopam ine-producing neurons
m ay be one of the m ost ag e-d ep en d en t neurons in the central
nervous system ( 1 0 ), it w ould not be surprising if a rise in prolactin
levels occurs in old age. This is clearly the fate for the senescent rat,
resulting in hyperprolactinem ic hypogonadotropic hypogonadism
(9). H owever, stim ulatory influences upon prolactin secretion clearly
also exist, an d rat studies has dem onstrated th at the p o sterio r
pituitary contains a potent prolactin-releasing factor (133). Thus, the
unchanged levels of serum prolactin between the 60- and 80-year-old
groups m ay be the net result of processes interfering w ith both
stim ulatory and inhibitory pathw ays of prolactin secretion. The
higher prolactin secretion rates in w om en has largely been attributed
to estrogen action (134). Estrogen adm inistration in rodents and
m onkeys increases the num ber of lactrotropes an d the rate of
prolactin synthesis (135). Despite declining levels of serum estrogen
in elderly w om en, the sex difference regarding basal prolactin and
especially TRH-stim ulated prolactin levels seems to persist in old
age. The reason for this is not clear.
No age influence upon basal serum AVP levels w as noted. This is
in line w ith reports of u n im p aired or even an enhanced AVP
response to osmotic stim uli w ith increasing age, at least in m en
(136,137). Sim ilar increm ents in AVP secretion after infusion of
hypertonic saline have also been found in young and old subjects,
w hereas ethanol ingestion leads to a late rise in AVP levels in
elderly stu d y subjects (136). In contrast to these findings, a low er
response in plasma AVP levels to an orthostatic challenge is seen in
older subjects (138). This m ay contribute to the augm ented AVP
response to hyperosmolality in the elderly.
Thus, despite certain morphological changes w ith increasing age
of the A V P-producing cells in the h y p o th a lam u s (58), AVP
production seems to be preserved in old age, or even enhanced to
stim uli such as dehydration. This m ay be of benefit to com pensate
for a reduced renal ability to conserve salt and w ater in aging man. It
m ay also com pensate for a reduced thirst that has been reported after
w ater deprivation in elderly men (137). The negative consequence
m ay be an increased risk in old age of developing excess AVP
39
secretion in response to exogenous stim u li such as certain
m edications, i.e., the syndrom e of in ap p ro p riate ADH secretion
(SIADH).
Paper II
The m in o r a g e -re la te d d eclin e in u r in a r y e x c re tio n of
norepinephrine (NE) in the present investigation contrasts w ith the
age-increased plasm a levels of this h o rm o n e th a t have been
repeatedly reported (22-24). The factor that explained m ost of the
variation in NE excretion in this study w as creatinine clearence,
irrespective of age. This is not surprising, considering that the
catecholamines excreted into urine are probably the result of both
filtration and tubular secretion (139). The un ch an g ed levels of
epinephrine (E) excretion w ith age how ever, correlates w ell w ith
unchanged plasma levels of the hormone w ith increasing age (22). In
a study of rat kidneys the existance of a renal E synthesizing enzyme
w as re c e n tly d e m o n stra te d (140). T h eo retically , th is m ay
counterbalance a decrease in filtration rate of E w ith increasing age.
The negative correlation between BMI and E excretion points tow ard
an increased activity of the sym pathoadrenal system in the elderly
(see also paper I).
There are several possible mechanisms by which dopam ine (DA)
excretion can decrease w ith age, such as decline in enzym e systems
in kidney tubuli, dietary changes including salt intake and changes
in plasm a levels secondary to a decreased secretion from nerve
terminals (30,31). The negative effect of age upon DA excretion is of
som e interest in relation to the rep o rted association betw een
sodium -sensitive hypertension and DA excretion levels (141). Thus,
a decline in DA levels has been postulated as a cause for the saltsensitive hypertension that has been proposed as the most common
type of hypertension in the elderly (142).
We discovered pronounced differences in the excretion of E and
NE in our elderly study group betw een smokers and non-smokers.
N icotine has w ide-spread and p ro n o u n ced effects on several
neurotransm itter systems and on pituitary horm one release (143).
This is m ediated through effects on cholinergic, m onoam inergic and
peptidergic systems in the brain (143). In acute studies of smoking,
40
elevated urinary catecholamine levels durin g sm oking have been
found (144) and cigarette sm oking w as related to high E an d NE
excretion rates in a com munity study of men (145) but not in another
stu d y of younger male volunteers (146). O ur results show that
differences betw een sm okers an d non-sm okers in E an d NE
excretion can persist u p to advanced age. Equal responses for young
and elderly men in plasm a E levels have also been reported during
psychological stress (147). This indicates a well preserved reactivity in
the sym pathoadrenal system in the aged.
Paper III and IV
Elevated activity in the hypothalam ic-pituitary-adrenal (HPA) axis in
stroke patients has been docum ented in these papers. A reduced
suppressibility by dexam ethasone (p ap er III) an d an increased
excretion of urine free cortisol (paper IV) were found.
Furtherm ore, disorientation w as associated w ith high activity in
the cortisol axis. M any CNS inform ation-processing capabilities are
supposed to be glucocorticoid-dependent (148) and based u p o n
anim al studies, glucocorticoids m ay exert a m odulatory influence
u po n behavior (149). Glucocorticoid receptors are present in rat
brain (5). In the rat, hippocam pal cells are im portant target sites for
corticosterone and express genes th at encode for tw o types of
corticosteroid receptors: the mineralocortiocoid (Type I) receptor and
the glucocortiocoid (Type II) receptor (150). The form er is involved
in the control of basal horm one levels thro u g h o u t the circadian
cycle, while the latter is concerned w ith the feedback control of the
stress response (151). The type II receptor also binds synthetic
glucocorticoids such as dexam ethasone. High glucocorticoid levels
have been advocated to be toxic to neurons (152). Therefore, high
cortisol levels could in d u ce a p ath o lo g ical change in the
hippocam pal neurons. This has also been verified in a series of
experiments by Sapolsky et al. (153). Hippocam pal activity norm ally
inhibits the cortisol axis, as dem onstrated by elctrical stim ulation in
m an (154). A dam age to hippocam pal neurons, and more subtly, loss
of hippocam pal corticosteroid receptors, can therefore be followed by
a d is tu rb a n c e of fe e d -b a c k r e g u la tio n w ith r e s u ltin g
hypercortisolism . In this note, that the hippocam pus seems to be
41
v ery sen sitiv e to ischem ic b rain d am ag e in rats, because
experim ental lesions cause im pairm ents of learning and m em ory
(155). Theoretically, hypercortisolism could reinforce the ischemic
dam age to hippocam pal neurons and thereby contribute to cognitive
disturbances after stroke.
Repeated stresses are common for stroke patients, such as various
cardiovascular com plications, infections, and em otional reactions.
These repeated stresses may increase the adrenal sensitivity to ACTH
(156), and therefore prolong the hypercortisolism . The prolonged
hypercortisolism m ight start a vicious cycle, in w hich dam age to
hippocam pal neurons may be a key point.
Low er b rain volum es w ere asso ciated w ith a d ecreased
su p p ressib ility to dexam ethasone. It has been suggested th a t
degenerative changes in the brain (e.g. in A lzheim er's disease) and
especially of the hippocam pus, can cause a m arked hyperactivity in
the cortisol axis and that hyperactivity could be more pronounced in
high age (157). In line w ith this, cognitive im pairm ent late after
stroke has also been shown to be associated w ith high cortisol levels
(63,65).
The cause of the correlation between frontally located lesions and
high cortisol axis activity is not obvious. However, the cortisol axis is
substantially influenced through the release of CRH from the
hypothalam us. Lesions in the frontal cortex in anim al experim ents
can have m arked effects on noradrenergic innervation (158). As NE
influences CRH release in anim al experim ents (156), anteriorly
located lesions m ay affect CRH release indirectly through effects
upon noradrenergic neurons. An increased activity of CRH may be
followed by a dow n-regulation of glucocorticoid receptors in the
hippocam pus, and this may induce a decreased suppressibility both
to endogenous corticosteroids and to dexam ethasone (153).
H yperactivity of the HPA axis is a com m on phenom enon in
major depression (159). Earlier studies of stroke patients have found
an association between the non-suppressive state of hypercortisolism
and major depression (63-65,70,71). Those studies included patients
at variable lengths of tim e after the initial stroke. We could not
verify such a linkage in our study group, despite a considerable
num ber of depressed stroke patients (31% of those w ho could be
evaluated). One reason for this may be that major depression can
42
have different etiologies at various time-points after stroke (160).
In paper IV, a clear association betw een m otor im pairm ent and
u rine free cortisol levels w as seen. V arious different stim ulating
influences upon the HPA axis m ay act at different locations, i.e.,
u p o n n eu ro tra n sm itto r in p u t to CRH secretion, u p o n ACTH
secretion and directly upon cortisol secretion. Therefore, it is possible
that the dexam ethasone suppression test and u rin ary free cortisol
levels m easure different influences on the HPA axis in acute stroke
patients.
A prolonged hypercortisolism can have a num ber of side-effects,
e.g. upon carbohydrate metabolism, upo n the m yocardium and the
im m une system (78,161). O n the o ther h and, there are m u tu al
connections betw een the HPA axis and the im m une system (162).
The com m on denom inator betw een body tem perature and cortisol
levels m ight be interleukin-1. This endogenous pyrogen seems to
have a key role in m any biologic processes. Interleukin-1 also
stimulates the release of CRF and ACTH (163) and has been proposed
as an im portant m essenger betw een the im m une system and the
HPA axis (162).
Prognostically, high plasm a cortisol levels have been associated
w ith a greater m ortality in stroke patients (62). The poorer functional
outcome associated w ith high cortisol excretion in paper IV is well in
line w ith this.
The co rrelatio n b etw ee n u rin e free c o rtiso l lev el a n d
catecholam ine excretion indicates a concom itant activation of the
HPA axis and the sym pathoadrenal system after stroke. This is not
surprising in view of the connections between the cortisol axis and
the sym pathoadrenal system at v arious levels, reinforcing one
another's activity (74,159).
Paper V
W hereas stroke patients as a group did not have an y o vert
abnorm alities in basal or stim ulated pituitary horm one levels, the
results indicate that severe stroke sym ptom s are associated w ith
such horm one disturbances. In addition, stroke m ay be associated
w ith changes in the levels of thyroid horm ones. The changes in
thyroid horm one levels are in line w ith other non-thyroid illnesses
43
w ith increased FT4 indices and slightly decreased FT 3 indices. The
m echanism s responsible for these changes are far from clear. The
m ost common reported abnorm ality is a depression of both total and
free serum T 3 levels (85). This has been interpreted as a decrease in
p eripheral conversion of T 4 to T 3 . The production of circulating
binding-inhibitor substances and a decrease in the concentration of
thyroxine-binding proteins m ay influence the levels of T 4 (85,164).
The net result for T 4 levels is variable, w ith increased levels in
m ilder forms of disease and often a gradual depression of T 4 (and T 3 )
levels in severe and long-standing disease states (165).
The m inor decrease in TSH response to TRH in stroke patients
contrasts som ew hat to an earlier study in w hich the TSH response
was significantly decreased am ong stroke patients (80). The patients
in that study, however, were studied at various tim e-points during
the hospitalization and included various types of cerebrovascular
disorders. On the other hand, TSH levels an d prolactin levels after
TRH w ere significantly associated w ith disorientation. A blunted
TSH response to TRH has been reported in patients w ith chronic
cognitive disturbances, i.e. A lzheim er's disease by some (166,167),
but not all (168,169) investigators. An excessive prolactin response to
TRH has also been found in A lzheim er's disease by some (168,170),
b ut not others (166,171).
A majority of our stroke patients showed a GH response to TRH
stim u latio n , i.e. they had a p aradoxical TRH response. This
response was associated w ith more extensive m otor im pairm ents. A
sim ilar finding has been rep o rted by M ashito et al. (81). The
m echanism for this response in patients w ith severe stroke is not
known. Based on anim al experim ents, hypothalam ic factors that
control GH secretion i.e. gro w th horm one releasing horm one
(GHRH) and som atostatin, seem to be extrem ely sensitive to
pharm acologic m anipulation of hypothalam ic n eu ro tran sm itto rs
(172). The release of neurotransm ittors from dam aged brain tissue
could therefore induce a change in the hypothalam ic regulation of
GH secretion.
The association betw een disorientation, m otor im pairm ent, level
of serum prolactin, GH and the pituitary-thyroid axis should be
investigated in fu rth er detail to explore if there is a com m on
44
suprah y p o p h y seal m echanism w hich regulate these p aram eters
a n d /o r if horm one changes may influence m ental performance.
Paper VI
This stu d y is the first long-term prospective investigation w ith
special em phasis on the prognosis for diabetic stroke patients.
Diabetes mellitus em erged as a major determ inant of m ortality after
stroke. W ith the use of C ox's prop o rtio n al h azard m odel, the
influence of diabetes u p o n m ortality could be stu d ied w hile the
increased proportion of cardiac disease that was seen in the diabetic
patients w as taken into account. The m ost pronounced influence
from diabetes upon m ortality was seen during the first m onths after
stroke. This is in agreem ent w ith m ost earlier studies dem onstrating
a w orse short-term prognosis after stroke (93-96,106). W ith our
m utivariate statistical approach we could also d em o n strate the
independent influence of diabetes on prognosis. This is in line w ith
the results of a recent American study (105), but not w ith the results
from a Sw edish study (104), both studies utilizing a m ultivariate
statistical technique.
The increased risk for recurrent stroke in diabetic patients has
been verified by others in recent studies (173,174). A m ore w ide­
spread large-vessel atherosclerosis in the cerebral circulation and in
the coronary vessels, as well as abnorm alities in vessel w alls and
blood com ponents m ay contribute to the increased risk for a
recurrent stroke and m yocardial infarction in diabetic patients
(92,175,176).
The diabetic p atien ts ten d ed to have a m ore p ro n u n ced
neurological deficit on adm ission to the hospital. Larger b rain
lesions in diabetic patients m ay be linked to the higher percentage of
cerebral embolism. Cerebral em bolism is associated w ith a m ore
pronounced neurological deficit th an non-em bolic stroke (177).
H igher hem atocrit levels in diabetic patients m ay contribute to this
increased cerebral throm boem bolic te n d en cy . A m o n g o th e r
abnorm alities an im paired autoregulation of cerebral blood flow
(178), a decreased deformability of erythro- and leucocytes (179,180),
hypercoagulability (181), hyperviscosity (182), a decreased synthesis of
prostacyclin w ith an increased adhesivity of throm bocytes (183), an
45
increased adhesion of erythrocytes to endothelial cells (184) and an
im pairm ent of the fibrinolytic system (185,186) have been described
in diabetic patients. The hyperglycemia per se m ay also contribute to
the developm ent of larger brain lesions, although this is a m atter of
controversy (115). In addition, our results indicate th at diabetic
patients seem to be prone to develop throm boembolic complications
after stroke, such as pulm onary embolism. This m ay partly be due to
the m ore pronounced neurologic deficit after stroke in these
patients.
Thus, diabetic patients represent a subgroup of stroke patients
w ho should be kept u n d e r especially close su p erv isio n . An
aggressive attitude against other risk factors for m acrovascular
disease seems pertinent in these patients.
GENERAL SUMMARY AND CONCLUSIONS
These investigations w ere u n d ertak en to stu d y the interactions
betw een different endocrine systems and stroke. Furtherm ore, two
studies w ere peform ed to establish endocrine param eters in those
age groups from which stroke patients are often recruited, i.e. 60 and
80-year old individuals were studied.
The results can be sum m arized as follows:
• Regarding hypophyseal horm ones in the elderly, prolactin and
AVP levels are unchanged w ith increasing age, w hereas the TSH
response to TRH declines som ew hat. N o change in p erip h eral
thyroid horm one levels could be detected. The activity of the cortisol
axis seems to be preserved, w ith unchanged plasm a and urine free
cortisol levels. This indicates a w ell p reserved activity in these
endocrine systems in advanced age, w ith the possible exception of a
decreased reserve capacity in the thyrotropin-producing cells in the
pituitary. Sex differences persist in old age, w ith higher levels in
w om en of basal and stim ulated serum prolactin levels and for
serum TSH response to TRH.
• Dopamine excretion is m arkedly reduced in high age. In contrast,
only slight decreases are seen for norepinephrine and epinephrine
excretion. The association b etw een d o p am in e ex cretio n an d
creatinine clearence underscores the im portant role for the kidneys
46
in dopam ine handling. The high levels of n o rep in ep h rin e and
epinephrine excretion in smokers points to a preserved activity in
o ld age of th e s tim u la to ry effects of sm o k in g o n th e
sym pathoadrenal system.
• Stroke subjects are exposed to hypercortisolism to a major degree
as manifested by a decreased suppressibility to dexam ethasone and by
an increased excretion of urine free cortisol. This hypercortisolism is
associated w ith disorientation, frontally located brain lesions, and
p o ssib ly also w ith m o to r im p a irm e n t a fte r s tro k e . The
dexam ethasone test has no value as an diagnostic aid for diagnosing
major depression in acute stroke patients. A concom itant activation
of the sym pathoadrenal system after stroke is su p p o rted by the
correlation betw een cortisol and norepinephrine excretion.
• In stroke patients, low th yrotropin levels an d high prolactin
levels after TRH stim ulation are associated w ith disorientation. A
paradoxical grow th horm one réponse to TRH is found in the
m ajority of stroke patien ts, an d high T R H -stim ulated g row th
horm one levels are associated w ith extensive pareses.
• Diabetic patients have a higher m ortality and are m ore often
affected by a recurrent stroke and m yocardial infarction after the
index event than non-diabetic patients.
47
ACKNOWLEDGEMENTS
This study was carried out at the D epartm ent of Internal M edicine,
Umeå U niversity Hospital.
I w ish to express m y gratitude to all those w ho have helped me
throughout this study, and especially to:
Professor Per Olov W ester, H ead of the D epartm ent of Internal
Medicine, for placing necessary facilities to my disposal and for his
continous interest in stroke unit care.
Associate Professors Erik Hägg and Kjell A splund, my supervisors,
for th eir w h o le-h earted in v o lv em en t an d excellent g u idance
throughout this study.
Dr. M atti V iitanen, co-w orker, for his scientific contrib u tio n s,
adm inistrative skills and introduction into the theories of survival.
Dr. Sture Eriksson, co-worker, for em inent statistical guidance.
D r.
M onica
Å strö m ,
co -w o rk er,
for
im p o rta n t
scien tific
contributions.
Drs. Åke Forssell, Kjell Grankvist and Yngve Gustafson, co-workers,
for good cooperation and fruitful scientific discussions.
Drs. Bo Carlberg, Folke Lithner, Bert-Ove Olovsson and Tage Strand
for their friendship and interest in scientific and clinical matters.
Dr. Pontus W iklund, my first tutor in Internal Medicine.
48
The stroke n urses M argareta Engde, M argot H elgesson, Eva
Sandström and Eva Kihl for their unlim ited helpfullness and care of
the stroke patients. H ead nurse Kristina Ö stm an and the rest of the
staff at our departm ent for their fruitful collaberation.
Ms. Ingegerd Söderström , Mrs. M argaretha D anielsson, Ms. Siri
M arakatt and Mr. Lars-Göran Sjöström for help in various aspects in
this study and for their never ending cheerfulness.
My family: Lillemor, Sara, A nton and A nna for their love and
support.
Last, but not least, I w ould like to express my sincere gratitude to the
patients and their relatives w ho m ade this study possible.
These studies were supported by grants from King G ustav V's 80th
A nniversary Fund, B orgerskapets i Umeå Research Foundation,
F und for M edical R esearch, U m eå U n iv ersity , 1987 Y ear's
Foundation for Stroke research, Stohnes Fund for Medical Research,
Joint Comm itte, N orth Region, Clas Groschinskys Foundation, the
H oechst Fund for D iabetes Research and the Sw edish M edical
Research Council (grant No. 27X-07192).
49
REFERENCES
1.
Martin JB, Reichlin S, eds. Clinical Neuroendocrinology. 2nd ed.
Philadelphia: FA Davis, 1987.
2.
Viitanen M, Fugl-Meyer KS, Bernspång B, Fugl-Meyer AR. Life
satisfaction in long-term survivors after stroke. Scand J
Rehab Med 1988;20:17-24.
3.
Parikh RM, Lipsey JR, Robinson RG, Price TR. Two-year
longitudinal study of post-stroke mood disorders: dynamic
changes in correlates of depression at one and two years.
4.
Stroke 1987;18:579-84.
Robinson RG, Bolla-Wilson K, Kaplan E, Lipsey JR, Price TR.
Depression influences intellectual impairment in stroke patients.
5.
Br J Psychiatry 1986;148:541-7.
McEwen BS, Weiss JM, Schwartz LS. Selective retention of
corticosterone by limbic structures in rat brain. N ature
6.
7.
1968;220:911-2.
Gregerman RI, Bierman EL. Aging and hormones. In Williams
RH, ed. W illiams' Textbook of Endocrinology. 6 th ed.
Philadelphia: W B Saunders, 1982:1192-1212.
Korenman SG. Endocrine Aspects of Aging. New York: Elsevier
Biomedical, 1982:1-7.
8.
Green MF. The endocrine system. In Pathy MSJ, ed. Principles
and Practice of Geriatric Medicine. Chichester: John Wiley &
Sons, 1985:914.
9.
Blackman MR. Aging. In DeGroot LJ, ed. Endocrinology. 2nd ed.
Philadelphia: W B Saunders, 1989:2348-66.
10.
Carlsson A. Neurotransmittors in old age and dementia. In
Häfner H, Moschel G, Sartorius N, eds. Mental Health in the
Elderly: A Review of the Present State of Research. Heidelberg:
Springer-Verlag, 1986:154-61.
11.
Morgan DG, Marcusson JO, Nyberg P et al. Divergent changes in
D-l and D-2 dopamine binding sites in human brain during
aging. Neurobiol Aging 1987;8:195-201.
50
12.
Pavlov EP, Harman SM, Chrousos GP, Loriaux DL, Blackman
MR. Responses of plasma adrenocorticotropin, cortisol, and
dehydroepiandrosterone
horm one
in healthy
to
aging
ovine
corticotropin-releasing
men. J Endrocrinol
Metab
1986;62:767-72.
13.
Blichert-Toft M. The adrenal glands in old age. In Greenblatt RB, ed.
Geriatric Endocrinology. New York: Raven Press, 1978:81-102.
14.
Parker L, Gral T, Perrigo V, Skowsky R. Decreased adrenal
androgen sensitivity to ACTH during aging. Metabolism
1981;30:601-4.
15.
Vermeulen A, Deslypere JP, Schelfhout W, Verdonck L, Rubens
R. Adrenocortical function in old age: response to acute
adrenocorticotropin stim ulation. J Clin Endocrinol Metab
1982;54:187-91.
16.
Selye H. Stress. Montreal: Acta, 1950:222-7.
17.
Munck A, Guyre PM, Holbrook NJ. Physiological functions of
glucocorticoids in stress and their relation to pharmacological
actions. Endocr Rev 1984;5:25-44.
18.
Stangl D, Pfohl B, Zimmerman M, Coryell W, Corenthal C
The relationship
19.
between
age
and
post-dexam ethasone
cortisol: a test of three hypotheses. J Affect Disord 1986;11:185-97.
Zimmerman M, Coryell W. Review: the dexamethasone
suppression test in healthy controls. Psychoneuroendocrinology
20.
1987;12:245-51.
Rosenbaum AH, Schatzberg AF, MacLaughlin RA et al. The
dexamethasone suppression test in normal control subjects:
comparison of two assays and effect of age. Am J Psychiatry
1984;141:1550-5.
21.
Lewis DA, Pfohl B, Schlechte J, Coryell W. Influence of
age on the cortisol response to dexamethasone. Psychiatry Res
1984;13:213-20.
22.
Morrow LA, Linares OA, Hill TJ, et al. Age differences in the
plasm a
cleareance
m echanism s
for
ep in ep h rin e
and
norepinephrine in humans. J Clin Endocrinol Metab 1987:65:508-11.
51
23.
Prinz PN, Halter J, Benedetti C, Raskind M. Circadian variation
of plasma catecholamines in young and old men: relation to
rapid eye movement and slow wave sleep. J Clin Endocrinol
Metab 1979;49:300-4.
24.
Pfeifer MA, Weinberg CR, Cook D, Best JD, Reenan A, Halter JB.
Differential changes of autonomic nervous system function
with age in man. Am J Med 1983;75:249-58.
25.
Wallin BG, Sundlöf G, Eriksson B-M, Dominiak P, Grobecker H,
Lindblad LE. Plasma noradrenaline correlates to sympathetic
muscle nerve activity in normotensive man. Acta Physiol Scand
26.
1981;111:69-73.
Kärki NT. The urinary excretion of noradrenaline and adrenaline
in different age groups, its diurnal variation and the effect of
muscular work on it. Acta Physiol Scand 1956; 39(Suppl
27.
132):l-96.
Brown MJ, Causon RC, Barnes VF, Brennan P, Barnes G,
Greenberg G. Urinary catecholamines in essential hypertension:
results of 24-hour urine catecholamine analyses from patients
in the medical research council trial for mild hypertension and
from matched controls. Q J Med 1985;57:637-51.
28.
Jenner DA, Harrison GA, Prior IAM, Leonetti DL, Fujimoto WY.
24-h Catecholamine excretion: relationships with age and
weight. Clin Chim Acta 1987;164:17-25.
29.
Bell C. Dopamine release from sympathetic nerve terminals.
Progr Neurobiol 1988;30:193-208.
30.
Suzuki H, Haruyuki N, Kawamura M, Yoshizawa M, Takeshita E,
Saruta T. Excretion and metabolism of dopa and dopamine by
isolated perfused rat kidney. Am J Physiol 1984;247:E285-E290.
31.
Lee MR. Dopamine and the kidney. Clin Sci 1982;62:439-48.
32.
Blumenthal HT, Perlstein IB: The aging thyroid. I. A description
of lesions and an analysis of their age and sex distribution. J
Am Geriatr Soc 1987;35:843-54.
52
33.
Bilous
RW,
T unbridge
WMG.
The
epidem iology
of
hypothyroidism-an update. Bailleries Clin Endocrinol Metab
34.
1988;2:531-40.
Robuschi G, Safran M, Braverman LE, Gnudi A, Roti E.
Hypothyroidism in the elderly. Endocr Rev 1987;8:142-51.
35.
Edén S, Sundbeck G, Lindstedt G et al. Serening for thyroid disease
in the elderly. Compr Gerontol [A] 1988;2:40-5.
36.
Tunbridge WMG, Brewis M, French JM et al. N atural
history of autoimmune thyroiditis. Br Med J 1981;282:258-62.
37.
Tzankoff SP, Norris AH. Effect of muscle mass decrease on agerelated BMR changes. J Appi Physiol 1977;43:1001-6.
38.
Gregerman RL, Gaffney GW, Shock NW. Thyroxine turnover in
euthyroid man with special reference to changes with age. J Clin
Invest 1962;41:2065-74.
39.
Spaulding SW. Age and the thyroid. Endocrinol Metab Clin
North Am 1987;16:1013-25.
40.
Snyder PJ, Utiger RD. Response to thyrotropin releasing
hormone (TRH) in normal man. J Clin Endocrinol Metab
1972;34:380-5.
41.
Snyder PJ, Utiger RD. Response to thyrotropin releasing
hormone in normal females over forty. J Clin Endocrinol Metab
1972;34:1096-8.
42.
Erfurth EM, Nordén NE, Hedner P, Hedner P, Nilsson A, Ek L.
Normal reference interval for thyrotropin response to thyroliberin:
dependence on age, sex, free thyroxin index, and basal
concentrations of thyrotropin. Clin Chem 1984;30:196-9.
43.
Jacques C, Schlienger JL, Kissel C, Kuntzmann F, Sapin R.
TRH-induced TSH and prolactin responses in the elderly. Age
Ageing 1987;16:181-8.
44.
Wenzel KW, Meinhold H, Herpich M, Adlkofer F, Schleusener H.
TRH-Stimulationstest mit alters- und geschlechtsabhängigem TSHAnstieg bei Normalpersonen. Klin Wochenschr 1974;52:722-7.
45.
Kaiser FE. Variability of response to thyroid-releasing
hormone in normal elderly. Age Ageing 1987;16:345-54.
53
46.
Ohara H, Kobayashi T, Shiraishi M, Wada T. Thyroid function of the
aged as viewed from the pituitary-thyroid system. Endocrinol
Jpn 1974;21:377-86.
47.
Harman SM, Wehmann RE, Blackman MR. Pituitary-thyroid
hormone economy in healthy aging men: basal indices of thyroid
function and thyrotropin responses to constant infusions of
thyrotropin releasing hormone. J Clin Endocrinol Metab
1984;58:320-6.
48.
49.
Yamaji T, Shimamoto K, Ishibashi M Kosaka K, Orimo H. Effects of
age and sex on circulating and pituitary prolactin levels in humans.
Acta Endocrinol 1976;83:711-8.
Sawin CT, Carlson HE, Geller A, Castelli WP, Bacharach P. Serum
prolactin and aging, basal values and changes with estrogen use and
50.
hypothyroidism. J Gerontol 1989;44:M131-5.
Blackman MR, Kowatch MA, Wehmann RE, Harman SM. Basal
serum prolactin levels and prolactin responses to constant
infusions of thyrotropin releasing hormone in healthy aging
men. J Gerontol 1986;41:699-705.
51.
Sun Y-K, Xi Y-P, Fenoglio CM et al. The effect of age
on number of pituitary cells immunoreactive to growth hormone
and prolactin. Hum Pathol 1984;15:169-80.
52.
Sonntag WE, Meites J. Decline in growth hormone secretion in
aging animals and man. Interdise Topics Geront 1988;24:11124.
53.
Ho KY, Evans WS, Blizzard RM et al. Effects of sex and age on the
24-hour profile of growth hormone secretion in man: importance
of endogenous estradiol concentrations. J Clin Endocrinol Metab
1987;64:51-8.
54.
Rudman D, Kutner MH, Rogers CM, Lubin MF, Fleming GA,
Bain RP. Impaired growth hormone secretion in the adult
population. J Clin Invest 1981;67:1361-9.
55.
Shibasaki T, Shizume K, Nakahara M et al. Age-related changes
in plasma growth hormone response to growth hormonereleasing factor in man. J Clin Endocrinol Metab 1984;58:212-4.
54
56.
Lang I, Schernthaner G, Pietschmann P, Kurz R, Stephenson JM,
Tempi H. Effects of sex and age on growth hormone response to
growth hormone-releasing hormone in healthy individuals. J
Clin Endocrinol Metab 1987;65:535-40.
57.
Pavlov EP, Harman SM, Merriam GR, Gelato MC, Blackman MR.
Responses of growth hormone (GH) and somatomedin-C to GHreleasing hormone in healthy aging men. J Clin Endocrinol
58.
Metab 1986;62:595-600.
Swaab DF, Fliers E, Hoogendijk JE. Vasopressin in relationship
to human aging and dementia. In Gash DM, Boer GJ, eds.
Vasopressin. New York: Plenum Press, 1987:611-25.
59.
Os I, Kjeldsen SE, Aakesson I et al. Evidence of age-related variation
in plasma vasopressin of normotensive men. Scand J Clin Lab
Invest 1985;45:263-8.
60.
Li CH, Hsieh SM, Nagai I. The response of plasma arginine
vasopressin to 14 h water deprivation in the elderly. Acta
Endocrinol 1984;105:314-7.
61.
Jenkins JS, Buckell M, Carter AB, Westlake S. Hypothalamicpituitary-adrenal function after subarchnoidal haemorrhage. Br
J Med 1969;4:707-9.
62.
Feibel JH, Hardy PM, Campbell RG, Goldstein MN, Joynt RJ.
Prognostic value of the stress response following stroke.
JAMA,-1977;238:1374-6.
63.
Lipsey JR, Robinson RG, Pearlson GD, Rao K, Price TR. The
dexamethasone suppression test and mood following stroke.
Am J Psychiatry 1985;142:318-23.
64.
Reding M, Orto L, Willensky P et al. The dexamethasone
suppression test. An indicator of depression in stroke but not a
predictor of rehabilitation outcome. Arch Neurol 1985;42:209-12.
65.
Agarwal A, Nag D, Kar AM, Agarwal AK, Chowdhury SR.
Dexamethasone suppression test & post stroke depression.
Indian J Med Res 1987;85:297-301.
55
66.
Charles GA, Lefevre A, Mirel J, Rush AJ. Suppression of
cortisol following dexamethasone in dem ented patients.
Psychiatry Res 1986;17:173-82.
67.
Carroll BJ, Feinberg M, Greden JF et al. A specific laboratory
test for the diagnosis of melancholia. Arch Gen Psychiatry
1981;38:15-22.
68.
Brown RP, Stoll PM, Stokes PE et al. Adrenocortical hyperactivity in
depression: effects of agitation, delusions, melancholia, and other
illness variables. Psychiatry Res 1988;23:167-78.
69.
70.
Dupont RM, Cullum CM, Jeste DV. Poststroke depression and
psychosis. Psychiatr Clin North Am 1988;11:133-49.
Finklestein S, Benowitz LI, Baldessarini RJ et al. Mood,
vegetative disturbance, and dexamethasone suppression test
after stroke. Ann Neurol 1982;12:463-8
71.
Bauer M, Gans JS, Harley JP, Cobb W. Dexamethasone
suppression test and depression in a rehabilitation setting.
Arch Phys Med Rehabil 1983;64:421-2.
72.
Bondy PK. Disorders of the adrenal cortex. In: Wilson JD, Foster
DW, eds. Textbook of Endocrinology 7th ed. Philadelphia: WB
73.
Saunders. 1985:816-90.
Meikle AW, Takiguchi H, Mizutani S, Tyler FH, West CD.
Urinary cortisol excretion determined by competitive proteinbinding radioassay: a test of adrenal cortical function. J Lab Clin
Med 1969;74:803-12.
74.
Goldstein DS. Stress-induced activation of the sympathetic
nervous system. Baillieres Clin Endocrinol Metab 1987;1:253278.
75.
Norris JW. Effects of cerebrovascular lesions on the heart.
Neurol Clin 1983;1:87-101.
76.
Tomamotsu T, Ueba Y, Matsomotu T et al. Electrocardiographic
observations and urinary excretion of catecholamines in
cerebrovascular accidents. Japn Circ J 1964;28:884-5.
56
77.
Jansen PAF, Thien T, Gribnau FWJ et al. Blood pressure and both
venous and urinary catecholamines after cerebral infarction. Clin
Neurol Neurosurg 1988;90:41-5.
78.
Cruickshank JM, Neil-Dwyer G, Stott AW. Possible role of
catecholamines, corticosteroids, and potassium in production of
electrocardiographic abnormalities associated with subarachnoid
haemorrhage. Br Heart J 1974;36:697-706.
79.
Delitala G, Tornasi P, Virdis R. Prolactin, growth hormone and
thyrotropin-thyroid hormone secretion during stress states in
man. Baillieres Clin Endocrinol Metab 1987;1:391-414.
80.
Hägg E, Asplund K, Eriksson S, Lithner F, Strand T, Wester P O.
Serum thyroid-stimulating hormone in cerebrovascular disease.
Acta Med Scand 1986;219:53-8.
81.
Mashito T, Takada I, Kenbo T, Watanabe A, Himuro K, Inagaki K.
Anterior pituitary function in patients with cerebrovascular
disease: with special reference to growth hormone secretion in
response to TRH administration. Folia Endocrinol Jap 1979;55:8995.
82.
Culebras A, Miller M. Dissociated patterns of nocturnal prolactin,
cortisol, and growth hormone secretion after stroke. Neurology
1984;34:631-4.
83.
Benvenga S, Morgante L, Bartalena L et al. Serum thyroid
hormones and thyroid hormone binding proteins in patients
with completed stroke. Ann Clin Res 1986;18:203-7
84.
Chopra IJ, Hershman JM, Pardridge WM, Nicoloff JT. Thyroid
function in nonthyroid illnesses. Ann Intern Med 1983;98:946-57.
85.
Tibaldi JM, Surks MI. Effects of nonthyroid illness on thyroid
function. Med Clin North Am 1985;69:899-911.
86.
Salonen JT, Puska P, Tuomilehto J, Homan K. Relation of
blood pressure, serum lipids, and smoking to the risk of
cerebral stroke. A longitudinal study in eastern Finland. Stroke
1982;13:327-33.
57
87.
Abbott RD, Donahue RP, MacMahon SW, Reed DM, Yano K.
Diabetes and the risk of stroke. The Honolulu Heart Program.
JAMA 1987;257:949-52.
88.
Davis PH, Dambrosia JM, Schoenberg BS et al. Risk factors
for ischemic stroke: a prospective study in Rochester,
Minnesota. Ann Neurol 1987;22:319-27.
89.
Barrett-Connor E, Khaw K-T. Diabetes mellitus: an independent
risk factor for stroke? Am J Epidemiol 1988;128:116-23.
90.
Boysen G, Nyboe J, Appleyard M et al. Stroke incidence and risk
factors for stroke in Copenhagen, Denmark. Stroke 1988;19:1345-53.
91.
WHO Study Group on Diabetes Mellitus. World Health
Organization Tech Rep Ser 727. Geneva:WHO, 1985.
92.
Helgason CM. Blood glucose and stroke. Stroke 1988;19:1049-53.
93.
Lithner F, Asplund K, Eriksson S, Hägg E, Strand T, Wester P-O.
Clinical characteristics in diabetic stroke patients. Diabete Metab
1988;14:15-9.
94.
Scmidt EV, Smirnov VE, Ryabova VS. Results of the
seven-year prospective study of stroke patients. Stroke
1988;19:942-9.
95.
Oppenheimer SM, Hoffbrand BI, Oswald GA, Yudkin JS.
Diabetes mellitus and early mortality from stroke. Br Med J
1985;291:1014-5.
96.
Terent A, Andersson B. The prognosis for patients with
cerebrovascular stroke and transient ischemic attacks. Ups J
Med Sci 1981;86:63-74.
97.
Wade DT, Skilbeck CE, Wood VA, Hewer RL. Long-term
survival after stroke, Age Ageing 1984;13:76-82.
98.
Marquardsen J. The natural history of acute cerebrovascular
disease. Acta Neurol Scand 1969;45:128-9.
99.
Abu-Zeid HAH, Choi NW, Hsu P-H, Maini KK. Prognostic factors
in the survival of 1,484 stroke cases observed for 30 to 48
months. II. Clinical variables and laboratory measurements.
Arch Neurol 1978;35:213-8.
58
100
Sheikh K, Brennan PJ, Meade TW, Smith DS, Goldenberg E.
Predictors of mortality and disability in stroke. J Epidemiol
Community Health 1983;37:70-4.
101
Kotila M. Declining incidence and mortality of stroke? Stroke
1984;15:255-9.
102.
Howard G, Walker MD, Becker C etal. Community hospital-based
stroke programs: North Carolina, Oregon, and New York. III. Factors
influencing survival after stroke: proportional hazards analysis of
103.
104.
4219 patients. Stroke 1986;17:294-9.
Chambers BR, Norris JW, Shurwell BL, Hachinski VC. Prognosis
of acute stroke. Neurology 1987;37:221-5.
von Arbin M. Atherosclerotic manifestations in stroke patients.
Clinical and epidemiological studies [Dissertation]. Stockholm:
Karolinska Institute, 1989.
105.
Howard G, Evans GW, Murros KE, Toole JF, Lefkowitz D, Truscott
BL. Cause specific mortality following cerebral infarction. J Clin
106.
Epidemiol 1989;42:45-51.
Asplund K, Hägg E, Helmers C, Lithner F, Strand T, Wester P-O.
The natural history of stroke in diabetic patients. Acta
Med Scand 1980;207:417-24.
107.
Pyörälä K, Laakso M, Uusitupa M. Diabetes and atherosclerosis:
an epidemiologic view. Diabetes Metab Rev 1987;3:463-524.
108.
Herlitz J, Malmberg K, Karlson BW, Rydén L, Hjalmarson Å.
Mortality and morbidity during a five-year follow-up of diabetics
with myocardial infarction. Acta Med Scand 1988;224:31-8.
109.
Cox DR. Regression models and life tables. J R Stat Soc
B 1972;34:187-220.
110.
Pulsinelli WA, Levy DE, Sigsbee B, Scherer P, Plum F. Increased
damage after ischemic stroke in patients with hyperglycemia with or
without established diabetes mellitus. Am J Med 1983;74:540-4.
111.
Siesjö BK. Cell damage in the brain: a speculative synthesis. J
Cereb Blood Flow Metab 1981;1:155-85.
112.
Berger L, Hakim AM. The association of hyperglycemia with
cerebral edema in stroke. Stroke 1986;17: 868-71.
59
113.
Ginsberg M, Welsh FA, Budd WW. Deleterious effect of
glucose pretreatm ent on recovery from diffuse cerebral
ischemia in the cat. I. Local cerebral blood flow and glucose
utilization. Stroke 1980;11:347-54.
114. Nedergaard M, Gjedde A, Diemer NH. Hyperglycemia protects
against neuronal injury around experimental brain infarcts.
Neurol Res 1987;9:241-4.
115.
Duverger D, MacKenzie ET. The quantification of cerebral
infarction following focal ischemia in the rat: influence of
strain, arterial pressure, blood glucose concentration, and age.
J Cereb Blood Flow Metab 1988;8:449-61.
116. Prado R, Ginsberg MD, Dietrich WD, Watson BD, Busto R.
Hyperglycemia increases infarct size in collaterally perfused
but not end-arterial vascular territories. J Cereb Blood Flow
Metab 1988;8:186-92.
117.
Strand T, Asplund K, Eriksson S, Hägg E, Lithner F, Wester PO. A
non-intensive stroke unit reduces functional disability and the
need for long-term hospitalization. Stroke 1985;16:29-34.
118.
Katz S, Akpom CA. A measure of prim ary sociobiological
119.
functions. Int J Health Serv 1976;6:493-508..
WHO Study Group on Diabetes Mellitus. World Health
Organization Tech Rep Ser 727. Geneva: WHO, 1985.
120. Joint international society and federation of cardiology/world
health organization (WHO) task force on standardization of
clinical nomenclature: nomenclature and criteria for diagnosis
121.
of ischemic heart disease. Circulation 1979;59:607-9.
Viitanen M, W inblad B, A splund K. A utopsy-verified
causes of death after stroke. Acta Med Scand 1987; 222:401-8.
122.
W alther WW. Simple, inexpensive micro method for the
estimation of T3 uptake. J Clin Pathol 1971;24:177-9.
123.
Gwee HM, Mashiter K. A 24 hour radioimmunoassay for human
prolactin. J Endocrinol 1978;77:423-4.
60
124.
Fyhrquist F, Wallenius M, Hollemans HJG. Radioimmunoassay
of vasopressin in unextracted plasma. Scand J Clin Lab Invest
125.
1976;36: 841-7.
Smedes F, Kraak JC, Poppe H. Simple and fast solvent
extraction system for selective and quantitative isolation of
adrenaline, noradrenaline and dopamine from plasma and urine.
J Chromatogr 1982;231:25-32.
126.
M atsui T, Hirano A. An atlas of the hum an brain for
computerized tomography. Tokyo: Igaku-Shoin Ltd, 1978.
127.
Haug G. Age and sex dependence of the size of normal
ventricles on computed tomography. Neuroradiology 1977;14:201-4.
128.
Wester P, Eriksson S, Forsell Å, Puu G, Adolfsson R. Monoamine
metabolite concentrations and Cholinesterase activities in
cerebrospinal fluid of progressive dementia patients: relation
to clinical parameters. Acta Neurol Scand 1988;77:12-21.
129.
Wilkinson L. SYSTAT: The System for Statistics. Evanston, IL:
SYSTAT Inc, 1988.
130.
Kaplan
EL, Meier
P.
N onparam etric
estim ation
from
incomplete observations. J Amer Statist Assn 1958;53:457-81.
131.
De Moor P, Steeno O, Meulepas E, Hendrikx A, Delaere K, Ostyn
M. Influence of body size and of sex on urinary corticoid
excretion in a group of normal young males and females.
Endocrinology 1963;23:677-83.
132.
Gregerman RI, Solomon N. Acceleration of thyroxine and
triiodothyronine turnover du rin g bacterial pulm onary
infections and fever: implications for the functional state of
the thyroid during stress and in senescence. J Clin
Endocrinol Metab 1967;27:93-105.
133.
Hyde JF, Ben-Jonathan N. The posterior pituitary contains a potent
prolactin-releasing factor: in vivo studies. Endocrinology
1989;125:736-41.
134.
Daughaday WH. The anterior pituitary. In Williams RH,
ed. Williams' Textbook of Endocrinology. 7th ed. Philadelphia:
W B Saunders, 1985:568-613
61
135.
MacLeod RM. Regulation of prolactin secretion. In Martin L and
Ganong WF, eds. Frontiers in Neuroendocrinology, voi 4.
136.
Philadelphia: Raven Press, 1975:169-94.
Helderman JH, Vestal RE, Rowe JW, Tobin JD, Andres R, Robertson
GL. The response of arginine vasopressin to intravenous ethanol
and hypertonic saline in man: the impact of aging. J Gerontol
1978;33:39-47.
137.
Phillips PA, Rolls BJ, Ledingham JGG et al. Reduced thirst
after water deprivation in healthy elderly men. N Engl J Med
1984;311:753-9.
138.
Rowe JW, Minaker KL, Sparrow D, Robertson GL. Age-related
failure of volum e-pressure-m ediated vasopressin release. J
Clin Endocrinol Metab 1982;54:661-4.
139.
Kopin IJ. Catecholamine metabolism: basic aspects and clinical
140.
significance. Pharm Rev 1985;37:333-64.
Ziegler MG, Kennedy B, Elayan H. Rat renal epinephrine
synthesis. J Clin Invest 1989;84:1130-3.
141.
Shikuma R, Yoshimura M, Kambara S et al. Dopaminergic
m odulation of salt sensitivity in patients with essential
hypertension. Life Sci 1986;38:915-21.
142.
Zemel MB, Sowers JR. Salt sensitivity and systemic hypertension
in the elderly. Am J Cardiol 1988;61:7H-12H.
143.
Clarke PBS. Nicotine and smoking: a perspective from animal
144.
studies. Psychopharmacology 1987;92:135-43.
W internitz WW, Quillen D. Acute hormonal response to
cigarette smoking. J Clin Pharmacol 1977;17:389-97.
145.
Reynolds V, Jenner DA, Palmer CD, Harrison GA. Catecholamine
excretion rates in relation to life-styles in the male population of
Otmoor, Oxfordshire. Ann Hum Biol 1981:8:197-209.
146.
Tucci JR, Sode J. Chronic cigarette smoking. Effect on
adrenocortical and sympathoadrenomedullary activity in man.
JAMA 1972;221:282-5.
62
147.
Barnes RF, Raskind M, Gumbrecht G, Halter JB. The effects of age on
the plasma catecholamine response to mental stress in man. J Clin
148.
Endocrinol Metab 1982;54:64-8.
Yates FE, Marsh DJ, Maran JW. The adrenal cortex. In Mountcastle
VB, ed. Medical Physiology. 14th ed. St Louis: CV Mosby, 1980,1558-
149.
1601.
Micco DJ Jr, McEwen BS. Glucocortiocoids, the hippocampus, and
behavior: interactive relation between task activation and steroid
150.
hormone binding specificity. J Comp Physiol Psychol 1980;94:624-33.
de Kloet ER, Reul JMHM. Feedback action and tonic influence of
corticosteroids on brain function: a concept arising from the
heterogeneity of brain receptor systems. Psychoneuroendocrinology
1987;12:83-105.
151.
Joels M, de Kloet ER. Effects of glucocorticoids
and
norepinephrine on the excitability in the hippocampus. Science
1989;245:1502-5.
152.
Sapolsky RM, Pulsinelli WA. Glucocorticoids potentiate ischemic
injury to neurons: therapeutic implications. Science,T985;229:13971400.
153.
Sapolsky RM, Krey LC, McEwen BS. The neuroendocrinology of
stress and aging: the glucocorticoid cascade hypothesis. Endocr
Rev 1986;7: 284-301.
154.
Mandell AJ, Chapman LF, Rand RW, Walter RD. Plasma
corticosteroids: changes in concentration after stimulation of
hippocampus and amygdala. Science 1963;139:1212.
155.
Auer RN, Jensen ML, Whishaw IQ. Neurobehavior al deficit due
to ischemic brain damage limited to half of the CAI sector of the
hippocampus. J Neurosci 1989;9:1641-7.
156.
Gaillard R-C, Al-Damluji S. Stress and the pituitary-adrenal axis.
Baillieres Clin Endocrinol Metab 1987;1:319-54.
157.
Sapolsky RM, McEwen BS. Why dexamethasone resistance? Two
possible
neuroendocrine
mechanisms.
In
Schatzberg
AF,
Nemeroff CB, eds. The Hypothalamic-Pituitary-Adrenal Axis.
New York: Raven Press, 1988:155-69.
63
158.
M orrison JH, Molliver ME, Grzanna R. N oradrenergic
innervation of cerebral cortex: widespread effects of local cortical
159.
lesions. Science 1979;205:313-6.
Gold PW, Goodwin FK, Chrousos GP. Clinical and biochemical
manifestations of depression. Relation to the neurobiology of
160.
stress. N Engl J Med 1988;319:413-20.
Robinson RG, Starr LB, Lipsey JR, Rao K, Price TR. A two-year
longitudinal study of post-stroke mood disorders: dynamic changes
in associated variables over the first six months of foilow-up. Stroke
1984;15:510-6.
161.
Buzzetti R, McLoughlin L, Scavo D, Rees LH. A critical assessment
of the interactions between the immune system and the
hypothalamo-pituitary-adrenal axis. J Endocrinol 1989;120:183-7.
162.
Bateman A, Singh A, Kral T, Solomon S. The immunehypothalam ic-pituitary adrenal axis. Endocr Rev 1989;10:92112 ..
163.
Beach JE, Smallridge RC, Kinzer CA, Bernton EW, Holaday JW, Fein
HG. Rapid release of multiple hormones from rat pituitaries
perfused with recombinant interleukin-1. Life Sci 1989;44:1-7.
164.
Oppenheimer JH, Schwartz HL, Mariash CN, Kaiser FE. Evidence for
a factor in the sera of patients with nonthyroidal disease which
inhbits iodothyronine binding by solid matrices, serum proteins,
165.
and rat hepatocytes. J Clin Endocrinol Metab 1982;54:757-66.
Nicoloff JT. Thyroid function in nonthyroidal disease.
In DeGroot LJ, ed. Endocrinology. 2nd ed. Philadelphia: W B
166.
Saunders, 1989:642-5.
Thomas DR, Hailwood R, Harris B, Williams PA, Scanlon MF,
John R. Thyroid status in senile dementia of the Alzheimer type
(SDAT). Acta Psychiatr Scand 1987;76:158-63.
167.
Sunderland T, Tariot PN, Mueller EA, Newhouse PA, Murphy
DL, Cohen RM. TRH stimulation test in dementia of the
Alzheimer type and elderly controls. Psychiatry Res 1985;16:269-75.
64
168. Peabody CA, Minkoff JR, Davies HD, Winograd CH, Yesavage J,
Tinklenberg JR. Thyrotropin-releasing hormone stimulation test
and Alzheimer's disease. Biol Psychiatry 1986;21:553-6.
169. Lampe T, Plymate SR, Risse SC, Kopeikin H, Cubberley L, Raskind
MA. TSH responses to two TRH doses in men with Alzheimer's
disease. Psychoneuroendocrinology 1988;13:245-54.
170. El Sobky A, El Shazly M, Darwish AK, Davies T, Griffin K,
Keshaven MS. Anterior pituitary response to thyrotrophin
releasing hormone in senile dementia (Alzheimer type) and
elderly normals. Acta Psychiatr Scand 1986;74:13-7.
171. Newhouse PA, Sunderland T, Tariot PN, Muller EA, Murphy DL,
Cohen RM: Prolactin response to TRH in Alzheimer's disease
and elderly controls. Biol Psychiatry 1986;21:963-7.
172.
Martin JB, Durand D, Gurd W, Faille G, Audet j, Brazeau P.
Neuropharmacological regulation of episodic growth hormone
and prolactin secretion in the rat. Endocrinology 1978; 102:106-13.
173.
Sacco RL, Foulkes MA, Mohr JP, Wolf PA, Hier DB, Price TR.
Determinants of early recurrence of cerebral infarction. Stroke
1989;20:983-9.
174.
Sobel E, Alter M, Davanipour Z, Friday G, McCoy R, Levitt LP,
Isack T. Stroke in the Lehigh Valley: combined risk factors for
recurrent ischemic stroke. Neurology 1989;39:669-72.
175.
Caplan LR, Gorelick PB, Hier D. Race, sex and occlusive
cerebrovascular disease. Stroke 1986;17:648-55.
176.
Freedman DS, Gruchow HW, Bamrah VS, Anderson AJ, Borboriak
JJ. Diabetes mellitus and angiographically-documented coronary
artery disease. J Clin Epidemiol 1988;41:659-68.
177.
Eriksson S, Asplund K, Hägg E, Lithner F, Strand T, Wester PO.
Clinical profile of cerebrovascular disorders in a population-based
patient sample. J Chron Dis 1987;40:1025-32.
178.
Bentsen N, Larsen B, Lassen NA. Chronically im paired
autoregulation of cerebral blood flow in long-term diabetics.
Stroke 1975;6:497-502.
65
179.
McMillan
DE,
U tterback
NG,
La
Pum a
J.
Reduced
erythrocyte deformability in diabetes. Diabetes 1978;27:895-901.
180.
Vermes I, Steinmetz ET, Zeyen LJJM, van der Veen EA.
Rheological properties of white blood cells are changed in diabetic
patients
181.
182.
with m icrovascular complications. Diabetologia
1987;30:434-6.
Small M, Lowe GDO, MacCuish AC, Forbes CD. Thrombin and
plasmin activity in diabetes mellitus and their association with
glycaemic control. Q J Med 1987;65:1025-31
Barnes AJ, Locke P, Scudder PR, Dormandy TL, Dormandy JA,
Slack J. Is hyperviscosity a treatable component of diabetic
183.
microcirculatory disease? Lancet 1977;ii:789-91.
Silberbauer K, Schernthaner G, Sinzinger H, Piza-Katzer H,
Winter M. Decreased vascular prostcyclin in juvenile-onset
184.
diabetes. N Engl J Med 1979;300:366-7.
Wautier J-L, Paton RC, Wautier M-P et al. Increased adhesion of
erythrocytes to endothelial cells in diabetes mellitus and its
relation to vascular complications. N Engl J Med 1981;305:237-42.
185. Gamba G, Solerte SB, Grignani G, Pacchiarini L, Montani N,
Ferrari E. Haemostatic variables, serum lipid abnormalities and
vascular complications in diabetes mellitus: a 5 year follow-up
study. Blut 1988;56:257-60.
186. Nilsson TK, Lithner F. Glycaemic control, smoking habits and
diabetes duration affect the extrinsic fibrinolytic system in type I
diabetic patients but microangiopathy does not. Acta Med scand
1988;224:123-9.
66