Lateralization of T-Lymphocyte Responses in Patients With Stroke
Effect of Sympathetic Dysfunction?
Background and Purpose A number of clinical observations indicate that stroke affects the course of immune-mediated diseases by lateralization of the disease manifestations, such as arthritis. The purpose of this study was to assess the impact of early stroke on lateralization of immune responsiveness.
Methods The delayed-type hypersensitivity (DTH) reaction to purified protein derivative was used as an in vivo measure of antigen-specific T-lymphocyte reactivity. Assessment of axon reflex vasodilatation was simultaneously used to test for cutaneous sympathetic activity.
Results There were no significant differences with regard to lateralization of DTH reactivity when all stroke patients were tested. However, patients with minor stroke displayed a significant (P<.001) decrease of DTH reaction on the paretic side compared with the contralateral side. In contrast, patients with major stroke showed a significant increase (P=.022) of DTH reaction on the paretic side. Patients with left hemiparesis had a significantly greater (P=.045) DTH response on the affected side than patients with a right hemiparesis. In addition, only the patients with motor deficit but not with sensory deficit or aphasia displayed side differences in DTH responses. When electrically evoked axon reflexes were studied in relation to DTH reactions, a significant correlation (r=.64; P<.001) was found between side asymmetries of DTH responses and side asymmetries of axon reflexes in an innervated skin area. No similar relation was present in skin areas where cutaneous sympathetic activity had been blocked by regional anesthesia.
Conclusions Early stroke lateralizes T-cell–mediated cutaneous inflammation. This effect depends on (1) the localization of the brain lesion, (2) the clinical course of the disease, and (3) the presence of motor deficit and may be mediated by (4) alteration of the cutaneous sympathetic nerve traffic.
A growing body of evidence suggests that the nervous system displays regulatory influences on the immune system.1 It has been demonstrated that lymphoid tissue in the spleen and lymph nodes is richly supplied by sympathetic nerve fibers.2 In addition, lymphocytes express receptors for many neuropeptides and neurotransmitters.3 Binding of pertinent neuropeptides to the receptors on the lymphocyte surface leads to a change of lymphocytic activity.3 Finally, it has been postulated that some regions of the brain may be crucial for the regulation of immune responses.4 5
Clinical reports have described that stroke affects the course of immune-mediated diseases, such as rheumatoid arthritis.6 Considering the relation between the nervous and the immune systems, it is tempting to speculate whether damage to the nervous system could influence immunologic status. In this respect, we recently demonstrated that established cerebral stroke with subsequent motor and sensory deficit significantly enhances antigen-specific T-cell reactivity in vivo on the paretic side of the body compared with the contralateral side.7 In addition, in vitro analysis of T-cell responses in patients with established stroke showed considerably higher numbers of responders to purified protein derivative (PPD) compared with a well-matched control group, indicating that stroke triggers systemic memory T cells.8
The aim of the present study was to evaluate the early effects of stroke on immune reactivity. The delayed-type hypersensitivity (DTH) test to PPD was chosen as a simple in vivo test of antigen-specific T-cell function. Peripheral blood mononuclear cells were studied with respect to phenotypic features and antigen-specific and mitogenic in vitro T-cell responses. In addition, a vascular reaction known to be influenced by sympathetic activity was related to the pattern of DTH responsiveness.
Subjects and Methods
Eighty subjects (41 men and 39 women, aged 18 to 96 years [mean±SD age, 68±15 years]), all patients at Sahlgrens’ University Hospital in Göteborg, were consecutively incorporated into the study. All the patients experienced stroke, which occurred between 0 and 60 days before DTH testing (mean±SD, 8.5±9.7 days). None of these patients had a history of previous stroke. Patients with malignant and autoimmune diseases or severe infections or on immunosuppressive drugs were excluded. All the patients were evaluated by a standardized examination of motor and sensory deficit, peripheral reflexes, muscular tone, and cranial nerve function. Occurrences of dysphasia and tactile neglect were also evaluated. The latter was tested by bilateral, simultaneous sensory stimulation (touch) of symmetrical parts of the body. Sixty-nine patients were examined with computed tomographic scan of the brain during the first days after the onset of stroke. Forty-four patients were reexamined 1 month later. In two cases, the first examination with computed tomographic scan was performed later than 1 month after the onset of stroke.
Stroke patients were classified according to Hachinski9 as having minor stroke, major stroke, or progressing stroke. The definition of minor stroke requires that the patient is discharged home, walks without assistance, and copes unaided with such self-care activity as eating, dressing, and toileting within 1 month after the disease onset.10 In contrast, patients with major stroke had stable and usually severe neurological deficit. Progressing stroke was defined as a deterioration of the neurological deficit between onset of the disease and the time of DTH test.
In vitro T- and B-cell responses as well as phenotypic patterns of lymphoid cells were analyzed prospectively in 10 patients (7 men and 3 women) with stroke. Heparanized blood samples were obtained for analysis on days 1, 2, 7, and 30 after the disease onset.
This study was approved by the ethics committee of the University of Göteborg.
Reagents and Procedures
Tuberculin and Histamine Tests
PPD (tuberculin) was purchased from Statens Seruminstitut and injected intradermally according to the manufacturer’s recommendations. Both arms of each subject were exposed to 2 transmission units of PPD, and the skin reaction was quantified after 72 hours. The test was judged positive if there was an induration of more than 5 mm in diameter. To study a non–T-cell–mediated induction of vasodilation with weal and flare reactions, 100 μL of histamine (0.1 mg/mL) was injected intradermally into the lateral aspect of both arms in each patient. The resulting response was recorded after 20 minutes. Two perpendicular diameters of the induration obtained were measured in both tuberculin and histamine tests.7 8 All tests were performed and quantified by the same person (E.T.).
Axon Reflex Vasodilation
Analysis of skin axon reflex vasodilation, an antidromic vasodilation evoked via polymodal afferent thin nerve fibers, was performed in 24 patients on the day of the DTH reaction. The test was performed on the dorsum of the second phalanx of the third and fourth fingers in both hands. Regional anesthesia of all four digital nerves was applied with 1 mL of 2% mepivacaine (Astra) at the base of the fourth finger of both hands to eliminate vascular effects of efferent sympathetic activity.11 The axon reflex was evoked by trains of 2-Hz constant-current impulses, 20 mA, 1 millisecond, 20 impulses, delivered by a circular metal electrode with a diameter of 6.8 mm, which served as a cathode. The reference electrode (the anode) was a silver plate covered by a cotton cloth soaked in physiological saline placed on the same arm, proximal to the wrist. Before and immediately after electric stimulation of the two fingers, skin perfusion was measured with a laser Doppler perfusion imager (Lisca). With this computer-controlled optical scanner a low-power He-Ne laser beam is moved step by step over the skin surface in a rectangular pattern. The scanning of an approximately 4×4-cm area (4096 sites) requires 3 to 4 minutes. In the presence of moving blood cells, a fraction of the light is Doppler-shifted, detected, and converted into an electric signal for further processing. The output signal relates linearly to tissue perfusion. The increase in skin perfusion after stimulation, which is mediated by polymodal C-fibers and attenuated by sympathetic influence,11 12 was calculated as the difference between the mean perfusion value of 30 measuring points at the stimulation site before and after electric stimulation (corresponding to the area of maximal perfusion increase). A detailed description of the technique has been recently published.12
In Vitro Stimulation of Peripheral Blood Lymphocytes (PBL) and Analysis of Interferon Gamma Production and Lymphocyte Phenotypes
The PBL were isolated by centrifugation on a Ficoll-Hypaque (Sigma Chemical Co) density gradient. The cells were used either directly for staining or resuspended at a concentration of 1×106/mL in 96-well round-bottomed microtiter plates (Nunc) in 0.2 mL Iscove’s medium (Gibco) supplemented with 10% heat-inactivated human AB serum, penicillin (100 U/mL), and 2.0 mmol/L l-glutamine. The PBL were stimulated with PPD (100 μg/mL), phytohemagglutinin (PHA) (5 μg/mL), and concanavalin A (ConA) (10 μg/mL) and cultured at 37°C in a moist atmosphere containing 5% CO2 in air. The cells were cultured for 72 hours, which has previously been found to be the optimal culture time for assessment of proliferative responses. During the final 8 hours of culture, 1 μCi of 3H-thymidine (Radiochemical Centre) was added to each well. The cultures were harvested into fiberglass filters, processed, and counted in a liquid scintillation counter. They were set up in triplicate, and the results are expressed as the mean of the counts per minute.
Levels of interferon gamma in supernatants from PPD-stimulated cell cultures were estimated by an enzyme-linked immunosorbent assay using monoclonal antibodies for coating and developing steps as previously validated.13
For lymphocyte phenotyping, PBL from the above 10 subjects were stained with monoclonal antibodies to CD3, CD4, CD8, HLA-DR, CD56, and CD25 (Becton-Dickinson). The cells were then subjected to two-color analysis by means of a FACStar (Becton-Dickinson).
Evaluation of B-Cell Reactivity
The ELISPOT assay (Czerkinsky et al,14 1983) was used to determine the respective number of immunoglobulin-secreting cells of the IgG, IgM, and IgA isotypes. Briefly, wells in the lids of 24-well culture cluster plates (Costar) were coated overnight at 4°C with 5 μg/mL of affinity-purified F(ab′)2 fragments of goat anti-human IgG (Cappel Laboratories), goat anti-human IgA, and goat anti-human IgM (Jackson Laboratories). The plates were then washed with tap water and blocked with 5% phosphate-buffered saline (PBS)/fetal calf serum for 1 hour. Next, 100 μL of mononuclear cell suspension was added into each coated well and incubated for 3.5 hours at 37°C. After another wash, biotinylated goat anti-human IgA, IgG, and IgM antibodies (Tago), diluted 1:750 in PBS containing 0.05% Tween 20 (polyoxyethylenesorbitan) (PBS-Tween) were applied, followed by incubation with 0.5 μg/mL of avidin-ALP (Sigma) in PBS-Tween. After addition of a phosphatase substrate solution (BCIP; Sigma), spots were enumerated under low magnification. Each spot represents one spot-forming cell. All the assays were done in duplicate and at cell concentrations ranging from 1×105 to 1×107 PBL per milliliter.
Statistical analysis regarding the differences between means was carried out by the Wilcoxon signed rank test. The χ2 test was used to analyze categorical data (Table 1⇓). Spearman’s rank order correlation method was used to calculate the coefficient of correlation and the level of significance between the axon reflex and the DTH reaction. A value of P<.05 was considered statistically significant.
Sixty of 80 patients included in the study displayed varying degrees of hemiplegia of the central type when examined at the time of DTH challenge. Forty-three of the hemiparetic patients were able to perform some voluntary movements, whereas 17 had a complete paralysis. Twenty patients displayed no motor deficit when tested for PPD. However, 5 of them had transient paresis initially, whereas 15 showed other stroke-related symptoms such as isolated hemisensory deficit or aphasia.
Twenty-seven patients displayed impaired sensory function. Among these patients, 17 had decreased proprioception, vibration, and pain sensation. Two patients showed loss of vibration only and 8 loss of pain only. Fifty patients had no sensory deficit at all. One patient displayed neglect symptoms, and 2 patients with paralysis could not collaborate in tests of sensory qualities because of a severe aphasia. Fifteen patients had aphasia: 2 without other symptoms, 7 associated with motor deficit only, and 6 with combined paresis and sensory loss. Forty patients had neurological signs on the right side of the body indicating a left-brain lesion, 39 on the left side indicating a right-brain lesion, and 1 on both sides indicating bilateral brain lesions.
Thirty-two of the 52 DTH-positive patients were classified as having a minor stroke and 19 as having a major stroke. These two groups of patients displayed significant differences in certain clinical and radiological features (Table 1⇑). At the time of testing, 1 patient had a progressive stroke. Within the population of DTH-positive probands, patients with minor stroke more frequently had diabetes mellitus (18% versus 0%), hypertension (46% versus 5%), and were on β-blocking therapy (34% versus 16%). Of the DTH nonresponders, 17 displayed minor stroke, whereas 11 suffered a major stroke. The DTH nonresponding stroke patients are excluded from all the subsequent statistical analyses.
Of the 10 patients chosen for the in vitro analysis, 9 displayed some degree of motor deficit (2 with paralysis, 7 with paresis), and 1 had aphasia and dyspraxia as the only symptom of stroke. Four patients also showed impaired sensory function on the paretic side, and 7 had some degree of aphasia. Eight had one or several symptoms from the cranial nerves.
DTH Reactivity in the Early Phase of Stroke
DTH Reactivity With Respect to Clinical Course of the Stroke
Fifty patients displayed a positive DTH reaction bilaterally, whereas 2 subjects showed a positive reaction only on the paretic side. Stroke patients were tested for DTH reactivity at different intervals after the onset of the disease (Fig 1⇓). Analysis of DTH responses with respect to disease duration revealed that approximately 20% (3 of 13) patients tested on days 0, 7, and 8 after stroke developed a DTH response. In contrast, approximately 70% of patients tested on days 1 through 6 (29 of 40) and on days 9 through 60 (20 of 27) showed DTH reactivity. All the 48 patients tested with intradermal injection of histamine exhibited a marked weal reaction on both arms, irrespective of the stroke duration. Twenty-eight stroke patients developed no DTH reaction at all. These patients were somewhat older than the DTH-positive group (73±16 versus 65±16 years), but otherwise there were only minor differences with regard to sex and major/minor stroke distribution.
There were no significant differences with regard to lateralization of DTH reactivity when all stroke patients were tested. However, patients with minor stroke (n=32) showed a significantly lower DTH reaction on the affected side of the body compared with the contralateral side (mean±SD difference, −2.7±7.2 mm; P<.001). In contrast, the patients with major stroke (n=19) showed a significantly stronger DTH reaction on the affected side (3.4±6.1 mm; P=.022) (Fig 2⇓). The histamine-induced weal did not display significant asymmetry in either patient group.
DTH Reactivity With Respect to Clinical Symptoms of the Stroke
Analysis of DTH reaction with regard to the type of neurological deficit showed that the 15 stroke patients without motor deficit when tested with PPD exhibited no differences in DTH reactivity between the sides of the body. However, in 5 of these patients who had an initial motor deficit (which totally regressed between the onset of stroke and the time of DTH challenge) the DTH reaction was significantly weaker (P<.005) on the affected side compared with the contralateral side (Table 2⇓). These 5 patients displayed all features of minor stroke. In contrast, the remaining 10 patients (8 with minor and 2 with major stroke) with stroke symptoms other than motor deficit exhibited no significant asymmetry of the DTH reaction (Table 2⇓).
Patients with paresis who exhibited a marked regression of motor deficit between the first examination at admission to hospital and the examination performed the day of DTH testing (n=16) showed a significant decrease (mean±SD difference, −4.8±8.3 mm; P= .007) of the DTH reaction on the affected side. In contrast, patients with stable motor deficit between the onset and the day of testing (n=17) had a significant increase (5.2±5.3 mm; P=.001) of the DTH reaction on the affected side (Fig 3⇓).
Patients with spasticity (n=21) (ie, increased reflexes and muscular tone) showed a nonsignificant increase of the DTH reaction between the two sides (2.4±6.9 mm; P=NS). However, this relation became significant (4.8±6.1 mm; P=.009) when only patients with major stroke (n=14) were taken into consideration. In contrast, patients without spasticity (n=30) displayed a decreased DTH reaction on the affected side compared with the nonaffected side (−3.7±15.4 mm; P<.001).
Lateralization of the Stroke-Induced Lesion in the Brain Affects DTH Reactivity
Twenty-six patients with positive DTH reaction had neurological symptoms on the right side and 24 on the left side. The number of patients with right and left hemiparesis, respectively, in relation to time of DTH testing and in relation to occurrence of major versus minor stroke was proportional. Patients with a left hemiparesis had a significantly greater DTH response on the paretic side compared with patients with a right hemiparesis (36.0±9.0 mm and 28.7±15.2 mm, respectively; P=.045). There were no significant differences in DTH reactivity between patients with minor and major stroke with regard to the side of hemiparesis.
Relation Between Antidromic Vasodilatation and DTH Reactivity
Twenty-four patients with stroke were simultaneously analyzed with respect to side asymmetries of axon reflex vasodilatation and DTH reactivity. When comparing nonanesthetized fingers, the pattern of the axon reflex vasodilation displayed a close correlation (r=.64; P<.001) to the pattern of the DTH reaction (Fig 4⇓, top panel). Thus, a more intense axon reflex vasodilation on the paretic side was related to a larger DTH reaction on the ipsilateral side and conversely. In contrast, when comparing the anesthetized fingers (Fig 4⇓, bottom panel), there was no correlation (r=.17; P=NS) between the side asymmetries of the axon reflex vasodilatation and the DTH reaction. There was no correlation between side asymmetries of the histamine weal and the axon reflex vasodilatation in the anesthetized and nonanesthetized fingers (data not shown).
Lymphocyte Reactivity In Vitro and Phenotypes of Peripheral Blood Mononuclear Cells During Stroke
To analyze the effect of the acute phase of the stroke on in vitro T-cell reactivity, responses of peripheral blood mononuclear cells to PPD, ConA, and PHA were studied in 10 patients on days 1, 2, 7, and 30 after the onset of symptoms. We observed decreased T-cell proliferative responses to PPD, PHA, and ConA during the first week after the onset of stroke (results not shown). In agreement with the above results, production of interferon gamma, a cytokine originating from activated T cells, in the supernatant from the lymphocytes stimulated by PPD was significantly decreased (P=.008) early during the stroke (Fig 5⇓).
The frequencies of immunoglobulin-producing B cells of the IgG, IgA, and IgM classes showed no significant variations between days 1, 2, 7, and 30 after the onset of stroke (results not shown).
Stroke did not alter the frequency of circulating CD3, CD4, CD8, CD3+DR+, CD3+25+, and CD3−56+ cells within 30 days after the onset (results not shown).
We have recently shown that established stroke enhances DTH reactivity on the affected side of the body.7 The present study deals with characterization of the early phase of stroke with respect to lateralization of T-cell–mediated inflammatory responses. Our study reveals four clinical features of stroke that might have directly or indirectly influenced the DTH response.
The clinical outcome of the stroke seems to be important for the lateralization of DTH responses. Thus, patients with minor stroke showed significantly weaker DTH reactivity on the affected side of the body, whereas those with major stroke showed a significantly greater DTH reactivity on the affected side than on the contralateral side. This somewhat puzzling finding may have occurred because of different consequences of minor and major stroke. In patients with minor stroke, a large part of the reversible neurological deficit was probably due to transient ischemia. Brain ischemia leads to local changes in pH, ionic balance, and increase of excitatory neurotransmitters, such as glutamate.15 16 Consequently, an increase of the excitability of the adjacent neurons with a subsequent inhibition of peripheral neurons, such as the lower motor neuron, would be expected to result in a decreased release of the neurotransmitter acetylcholine and its colocalized neuropeptide calcitonin gene-related peptide.17 Since both these substances act as T-cell activators,18 19 minor stroke would lead to downregulation of DTH reactivity on the affected side of the body. In contrast, major stroke with a permanent neurological deficit from the very onset of the disease caused by a localized neuronal death would lead to a loss of neuronal function, including central inhibitory properties. Resulting increased activity of peripheral neurons might trigger increased secretion of acetylcholine and calcitonin gene-related peptide, leading to an increase of DTH reactivity on the affected side of the body. Support for the organic differences between groups of patients with minor and major stroke is provided by a computed tomographic scan analysis indicating that stroke patients with stationary symptoms had higher frequency of visualized brain lesions (compare Table 1⇑).
The second clinical feature related to asymmetrical DTH responses is the presence of motor deficit. Patients with intact motor function did not exhibit any differences in DTH reactivity between the paretic side and the contralateral side. This is in accordance with our previous study demonstrating that motor deficit was required to give rise to increased DTH reactivity on the paretic side in patients with established stroke.7 Even here a changed pattern of neurotransmitter and/or neuropeptide release could have influenced the DTH reaction in these patients. This hypothesis is supported by the demonstration of a significant relation between presence of spasticity, a clinical sign of downregulation of central motor control, and an altered DTH reaction on the paretic side compared with the contralateral side.
Hemiparesis-induced immobilization of the limb might have potentially contributed to the lateralization of DTH reactivity by, for example, decreasing blood and lymphatic flow, thereby downregulating vasodilatory capacity and consequently the inflammatory response. However, cutaneous response to histamine, a probe of vasodilatory capacity, was not affected by the paresis. In addition, a previous study has demonstrated that phenolsulfonphthalein clearance time from the skin is similar in normal and hemiplegic limbs.20
The third clinical feature of stroke associated with changes of DTH reactivity is the localization of the brain lesion. Patients with a lesion in the right brain hemisphere had a significantly larger DTH reaction on the paretic side than patients with a left-sided lesion. This could not be explained by left hemisphere lesions being dominated by minor stroke or a preponderance of major strokes among right hemisphere lesions. Our results point rather to the possibility that inhibitory and stimulatory centra for modulation of immune responses exist in the right and left hemispheres, respectively. In this regard, Renoux et al21 and Neveu22 have demonstrated that experimental lesions in the frontal, parietal, or occipital cortices of mice on the left side depress systemic T-cell responses, whereas similar lesions in the right cerebral cortex have an opposite effect.
Finally, we have demonstrated that asymmetries of the cutaneous DTH responses correlated with asymmetries of cutaneous axon reflex vasodilatation. This correlation was, however, present only in nonanesthetized fingers. Since the anesthesia blocks efferent sympathetic nerve traffic, the magnitude of the axon reflex in innervated skin is influenced by sympathetic vasoconstrictor activity at the small-vessel level. This agrees with previous results demonstrating that axon reflex vasodilatation is reduced by increases of skin vasoconstrictor (but not sudomotor) nerve traffic.10 Thus, the present findings suggest that an asymmetry of cutaneous sympathetic vasoconstrictor activity accounts for the asymmetry of the axon reflex vasodilatation.
Stroke-induced brain lesions followed by asymmetries of peripheral sympathetic outflow have recently been demonstrated in an animal model.23 Interestingly, we have recently noted that in one patient with stroke leading to hemiparesis there was also a significant asymmetry of resting sympathetic nerve traffic to the leg muscles (H.N., MD, and B.G.W., MD, unpublished data).
Asymmetries of sympathetic traffic may influence the DTH response in two ways. Altered sympathetic activity would lead to changes of blood flow,24 which could influence the inflammatory response measured by DTH reaction. The parallel changes in DTH reactivity and histamine-induced weal might support this hypothesis since vasodilation is a common condition for these two reactions.25 However, in contrast to the strong correlation between asymmetries of DTH responses and axon reflexes, we found no significant correlation between asymmetries of the histamine-induced weal and axon reflexes. This suggests that another mechanism, such as a direct effect of sympathetic transmitters on inflammatory cells, may mediate the putative effects of the sympathetic nervous system on DTH responses. This hypothesis is supported by the studies of Felsner et al,26 demonstrating that norepinephrine, the main sympathetic neurotransmitter, inhibits T-cell function. Decrease of proliferative responses and interferon gamma production in vitro in response to PPD early during the stroke further corroborates the in vivo data. In vivo anergy to PPD, observed early after the onset of the stroke, indicates (in addition to lateralization) systemic changes of immune reactivity as a consequence of a brain lesion. These systemic changes of T-cell reactivity during stroke might have been due to increased sympathetic activity and/or hypercortisolism reported to occur early after the stroke.27 28
In conclusion, our results show that early stroke influences T-cell responses. This effect, although seen systemically, is predominantly lateralized and closely related to (1) side of the brain lesion, (2) clinical outcome of the disease, (3) motor deficit, and (4) alterations of sympathetic activity. The limitations in our study include the clinical heterogeneity of the stroke patient population and hence the need for subgroup statistical analysis. A future prospective study with a more sizeable patient population is presently planned.
This study was supported by grants from the Göteborg Medical Society, University of Göteborg, the Swedish Association Against Rheumatism, the King Gustaf V’s 80-Year Foundation, the Stroke Foundation, the Gamla Tjänarinnor Foundation, and the Swedish Medical Research Council. We thank Lena Svensson and Margareta Verdrengh for excellent technical assistance.
- Received May 16, 1994.
- Revision received August 22, 1994.
- Accepted October 3, 1994.
- Copyright © 1995 by American Heart Association
Ackerman KD, Felten SY, Bellinger DL, Livnat S, Felten DL. Noradrenergic sympathetic innervation of spleen and lymph nodes in relation to specific cellular compartments. Prog Immunol. 1987;6:588-600.
Carr DJJ, Blalock JE. Neuropeptide hormones and receptors common to the immune and neuroendocrine systems: bidirectional pathway of intersystem communication. In: Ader R, Felten DL, Cohen N, eds. Psychoneuro-immunology. 2nd ed. San Diego, Calif: Academic Press, Inc; 1991:573-588.
Besedovsky HO, Del Ray AE, Sorkin E. Immuno-neuroendocrine interactions. J Immunol. 1985;135:750-754.
Thompson M, Bywaters EGL. Unilateral arthritis following hemiplegia. Ann Rheum Dis. 1962;21:370-377.
Hachinski V. Classification of stroke for clinical trials. Stroke. 1990;21:27-29.
Blomstrand C, Olsson JE, Nilsson B, von Arbin M, Britton M, Elwin CE, Elmers C, Norrving B, Rosen A, Samuelsson K, Strandberg K, Wahlgren NG. Swedish Aspirin Low-Dose Trial (SALT) of 75 mg aspirin as secondary prophylaxis after cerebrovascular ischaemic events. Lancet. 1991;338:1345-1349.
Cooper P, Martin J. Neuropeptide functions. In: Asbury AK, McKhan GM, McDonald WI, eds. Diseases of the Nervous System. Philadelphia, Pa: WB Saunders; 1986;1:73.
Bevan JA, Bevan RD, Duckles SP. Adrenergic regulation of vascular smooth muscles. In: Bohr DF, Somlyo AP, Sparks HV, eds. Handbook of Physiology, Section 2: The Cardiovascular System, Volume II. Bethesda, Md: American Physiological Society; 1980:515-566.
Hammarlund A, Olsson P, Pipkorn U. Blood flow in dermal allergen-induced immediate and late phase reactions. Clin Exp Allergy. 1988;19:197-202.
Olsson T, Marklund N, Gustafson Y, Näsman B. Abnormalities at different levels of the hypothalamic-pituitary-adrenocortical axis early after stroke. Stroke. 1992;23:1573-1576.