J Neural Transm [P-D Sect] (1992) 4:53-68
__ Journal o f Neural Transmission 9 Springer-Verlag 1992 Printed in Austria
Neurochemical markers in the cerebrospinal fluid of patients with Alzheimer's disease, Parkinson's disease and amyotrophic lateral sclerosis and normal controls P. Hartikainen 1, K. J. Reinikainen 1, H. Soininen 1, J. Sirvifi2, R. Soikkeli 1, and P. J. Riekkinen 1
1Department of Neurology, and 2Neurochemical Research Laboratory, University of Kuopio, Finland Accepted May 27, 1991
Summary. Several neurotransmitter markers were investigated in the cerebrospinal fluid (CSF) from patients with Alzheimer's disease (AD) (n= 27), Parkinson's disease (PD) (n = 35) and ALS (n=26) and from control subjects (n-- 34) to compare the possible alterations in the biochemical profiles of these different neurodegenerative diseases. The main proportion of the patients represented an early phase of the illness at the time of the diagnosis. Correlations of the degree of dementia and the stage of the disease with CSF measures were evaluated. The CSF levels of somatostatin like-immunoreactivity (SLI) were significantly reduced in AD patients when compared with those of normals and ALS patients. The CSF concentrations of homovanillic acid (HVA) were significantly decreased for PD patients and the decrease focused on the nondemented patients. A trend of decreasing HVA values towards the most advanced stage of Parkinson's disease assessed by Webster's scale was also displayed. The content of 3-methoxy-4-hydroxyphenylglycol (MHPG) in the CSF was higher for ALS patients than for other groups. The lowest 5-hydroxyindoleacetic acid (5HIAA) levels were observed in the PD group and the lowest acetylcholinesterase (ACHE) activities were found in the PD patients with the most severe disease. Changes in CSF measures were too subtle to be beneficial for diagnostic purposes, but adequate for reflecting the different neurochemical profiles of these three degenerative neurological disorders.
Keywords: Cerebrospinal fluid, neurotransmitters, Alzheimer's disease, Parkinson's disease, ALS.
54
P. Hartikainen et al. Introduction
Alzheimer's disease (AD), Parkinson's disease (PD), and amyotrophic lateral sclerosis (ALS) are progressive neurological degenerative disorders of humans in middle and late life. Typically, clinical signs and symptoms are different, although the diseases share some similar features; patients with AD show dementia and later often extrapyramidal signs, PD patients develop a motor disorder with or without dementia and patients with ALS show devasting and weakness of musculature, while dementia in this disease is rare. Neurochemical studies of these diseases have indicated the complexity of the neurotransmitter systems involved. Post mortem studies have consistently shown the deficit of the ascending cholinergic system in AD and the disruption of the nigrostriatal pathways in PD (Rossor et al., 1982; Perry, 1986; Ruberg et al., 1986). The involvement of the ascending cholinergic innervation has also been found in demented PD patients (Dubois et al., 1983). Disturbances of noradrenergic and serotonergic systems have been described both in AD and PD (Rossor and. Mountjoy, 1986; Palmer et al., 1987). Data concerning neurochemical alterations in ALS are limited and difficult to interpret due to widely distributed pathological changes both in the spinal cord and the brain. A proportion of the changes in these neurotransmitter markers reflects to the cerebrospinal fluid (CSF). The most unequivocal finding is a low concentration of homovanillic acid (HVA), a metabolite of dopamine, in the CSF of PD patients (Johansson and Roos, 1967). Many studies have detected reduced cholinesterase (ChE) levels in the CSF from AD and demented PD patients, but not consistently (see Discussion). Therefore, the value of neurotransmitter measurements of the CSF as a diagnostic tool is doubtful. The present study attempts to clarify the profile of several transmitter related parameters in the CSF from patients with AD, PD, and ALS and from controls in relation to the degree of dementia, the stage of the disease and some clinical manifestations. The patients are carefully investigated using the standardized protocol of clinical, neuropsychological, neurophysiological and CSF examinations. Patients and methods Patients
The study included 128 subjects who were admitted to the Neurological Department of Kuopio University Central Hospital for diagnostic or therapeutic evaluation. The subject or, when needed, her/his nearest relative gave informed consent for participation in the study. The study was approved by the local ethics committee. All subjects passed through the following study battery: general physical and neurological examination, assessment of Hachinski Ishaemic score (Rosen et al., 1980), a measurement of blood pressure, ECG, chest X-ray, EEG, spectral analysis of EEG, somatosensoric evoked potentials, visual evoked potentials, brain CT-scan and neuropsychological tests. Laboratory tests consisted of blood cell count, erythrocyte sedimentation rate, blood glucose, serum urea, electrolytes,calcium, phosphate, vitamin B12, erythrocyte folate, liver
Neurochemical markers in the cerebrospinal fluid
55
and thyroid function tests, cholesterol, triglycerides, immunoglobulin (IgG), albumin (alb), lues serology and urine analysis. The routine CSF analysis comprised count of erythrocytes and leukocytes, determination of the total protein content, protein electrophoresis, IgG, alb, IgG/alb ratio, IgG index = (CSF IgG/alb)/(serum IgG/alb). The control population comprised 34 subjects (20 females, 14 males; mean age 49.3 + 16.7 years, range 20-80 years). The subjects had no history of dementia, chronic psychiatric disease or depression and did not use any psychotropic drugs. The subjects were selected from patients investigated due to minor neurological symptoms. In all cases the above mentioned tests were normal. The Alzheimer group consisted of 33 patients (24 females, 9 males; mean age 72.1 + 7.5 years, range 55-85 years) who fulfilled the diagnostic criteria of probable AD according to the definition of the NINCDS-ADRDA Work group (McKhann et al., 1984). All but two patients who were presently diagnosed were at diagnostic investigations. Metabolic, endocrine, nutritional and infectious causes of dementia were excluded. The patients with evidence of cerebrovascular disease according to the modified ischemic scale of Hachinski were excluded. For comparison of CSF findings, we divided the patients into presenile (age < 65 years) and senile dementia groups (age >~ 65 years) according to the onset of the disease and into three groups of clinical severity assessed by the Clinical Dementia Rating scale (CDR) (Hughes et al., 1982) mild (score 0.5-1), moderate (score 2) and severe (score 3) dementia. Extrapyramidal signs and symptoms were evaluated by Webster's scale (Webster, 1968). Five of the patients with severe dementia were receiving neuroleptics, four moderately or severely demented patients and one mildly demented patient were on codergocrin mesylate regimen. One patient with severe dementia had benzodiazepine medication. The duration of the disease from the first signs of the disease was less than 1 year in 2 patients, 1-2 years in 8 patients, 2-5 years in 21 patients, and over 5 years in two patients. Idiopathic PD patients were 12 females and 23 males with mean age 63.9 + 10.4, range 40-80 years. Twenty-eight patients were in diagnostic examinations, 7 patients represented earlier diagnosed and more advanced disease. The stage of the illness was evaluated according to the criteria of Hoehn and Yahr (1967) and clinical signs and symptoms were assessed by Webster's scale (Webster, 1968). The duration of PD was up to 1 year in 9 patients, 1-2 years in 12 patients, 2-5 years in 4 patients, and over 5 years in 7 patients. The severity of dementia was assessed by the CDR scale as in the AD group. Of 10 demented cases (28.6% from all PD patients) 7 patients showed mild and 3 moderate dementia. Eleven patients were on continuous antiparkinsonian treatment; 7 patients received L-dopa, 4 deprenyl, 1 bromocriptine, 1 amantadine, 3 neuroleptics, and 1 patient was on benzodiazepine regimen. The regimen combinations were variable. The patients with L-dopa represented stages IV-V of Hoehn and Yahr scale and most of the other medicated patients were of milder stages of the disease. Twenty-six patients with ALS (8 females, 18 males; mean age 61.9 + 11.2 years, range 28-76 years) were selected for this study. The diagnosis was based on history and clinical examination and it was confirmed b y electromyography showing typical denervation changes, fibrillations, fasciculations and loss of motor units in the peripheral musculature. Furthermore, muscle biopsy and serum creatine kinase values were investigated. Seven patients had predominantly bulbar symptoms and signs, 8 patients represented mainly the spinal form and 11 patients the mixed type of the disease. All ALS patients were at the early stage of the disease in the diagnostic investigations. ALS patients were estimated by the CDR scale: Five of the ALS patients scored 0.5 (questionable dementia), five scored 1 (mild dementia), and all other presented the score 0 (no dementia). In addition, cognitive abilities of ALS patients were assessed by using the Wechsler Adult Intelligence Scale (WAIS) performance IQ scores (PIO). The PIQ score of clinically possible ALS-dementia
56
P. Hartikainen et al.
cases (75.6 + 14.2) differed significantly (p < 0.001) from the scores of non-demented patients (98.6 • 17.9) and age-matched controls (100.8 + 9.1).
CSF samples The subjects were kept on low monoamine diet before the lumbar puncture. The CSF samples were collected after an overnight bed rest and fasting at 08.00-10.00 a.m. the subjects lying in a lateral decubitus position. The medicated patients were drugfree for at least 4 days before the lumbar puncture. The first specimen of CSF, 4 ml, was collected in 1 ml fractions for the routine analysis and the next specimens, together 15 ml in 5 ml fractions were collected for other determinations. Specimens were immediately frozen in liquid nitrogen and stored at - 80 ~ until assayed. Levels of neuropeptides, somatostatin (SLI), adrenocorticotropic hormone (ACTH) and f3-endorphin (BLI), were determined from the first fraction. The activity of acetylcholinesterase (ACHE), concentrations of monoamine metabolites; homovanillic acid (HVA), 5-hydroxyindoleacetic acid (5HIAA), and 3-methoxy-4-hydroxyphenylglycol(MHPG), and protein were determined from the second fraction. The activity of another noradrenergic marker, dopamine-~-hydroxylase (DBH), was measured from the third fraction.
Chemical methods The total protein content of the CSF was measured by the method of Lowry et al. (1951). The levels of monoamine metabolites were determined by using high performance liquid chromatography (HPLC) with electrochemical detection as described previously (Jolkkonen et al., 1987 a). The activity of DBH in the CSF was detected by a radioenzymatic method as reported by Soininen et al. (1984 a). The activity of AChE was determined as described previously by Sirvi6 et al. (1989a). Levels of SLI, BLI, and ACTH immunoreactivity in CSF were measured by radioimmunological assays (Jolkkonen et al., 1987 b, 1990).
Statistical methods The normal distribution of the analysed variables was tested with Kolmogorov-Smirnov test. Differences between groups were evaluated by analysis of variance (ANOVA) adjusted for confounding factors (ANCOVA), or by the non-parametric Kruskall-Wallis test when the assumptions for ANOVA, particularly the equivalency of variance or the normality of the distribution did not fulfil. The post hoc analyses were consequently Student's t-test with Bonferroni correction when appropriate or a modified procedure developed for Kruskall-Wallis analysis (Siegel and Castellan, 1988). In addition, when applying KruskallWallis analysis, the influence of confounding factors was checked by calculating Pearson correlation coefficients for CSF measures. The relationships between clinical measures (Webster's scale and Hoehn & Yahr scale) and CSF variables were assessed by using the non-parametric Spearman correlation test.
Results The m e a n CSF levels o f n e u r o t r a n s m i t t e r related variables o f patients with A D , P D a n d A L S a n d controls are shown in Table 1. Because age, height, weight, a n d season were f o u n d to interfere with the levels C S F measures in our previous w o r k with controls ( H a r t i k a i n e n et al., 1991), all C S F variables but M H P G , D B H , a n d BLI were adjusted for age and, in addition, H V A for height a n d season, A C T H for season a n d weight a n d SLI for weight. The c o n c e n t r a t i o n o f M H P G in the C S F correlated positively with age in n o r m a l subjects (Har-
Neurochemical markers in the cerebrospinal fluid
57
tikainen et al., 1991) and in A D patients (r -- 0.40, p = 0.04), but not in PD (r --- 0.09, p = 0.61) or ALS patients (r = 0.29, p --- 0.10). Seasonal alterations in the CSF levels of DBH and BLI found for normals (Hartikainen et al., 1991) were not observed for any of the disease groups. A C T H values for AD patients (p < 0.05) showed similar seasonal change as did those of controls; the levels were higher in the light than in the dark period (Hartikainen et al., 1991). The total protein content in the CSF showed great interindividual variability in each group. PD patients and ALS patients differed significantly from controls (Table 1). The activity of AChE in the CSF was lowest for A D patients, but the differences did not reach statistical significance between the groups (Table 1). In PD, the activity of AChE in the CSF was reduced in patients with the most advanced disease (Table 4). The AChE activities of demented PD patients were lower than those of non-demented patients (Table 3). The ALS patients had the highest M H P G concentrations of the CSF, and the ALS group differed significantly from the other groups (Table 1). The M H P G levels for demented and non-demented ALS patients were comparable as were all other CSF measures but SLI (Table 5) which, however, did not differ from those of controls (Table 1). Neither did the type of ALS contribute to any changes in neurotransmitter markers of the CSF (data not shown). The DBH activities of the CSF did not differ across the groups (Table 1). As expected, the CSF H V A levels were lowest for PD patients and differed significantly from all other groups (Table 1). In line with previous observations, height, as a covariate in the analysis of variance (ANOVA), showed a significant effect on H V A values of the CSF (F = 19.6, p = 0.0001). The concentrations of 5HIAA were lowest for the PD patients and differed significantly from those of AD patients, but not from the controls (Table 1). SLI values of the CSF were lowest for the AD group and differed significantly from the controls (Table 1). The BLI levels of the CSF did not differ across the groups (Table 1). Neither did the A C T H values of the CSF show any differences between groups (Table 1). In comparison of patients with presenile or senile dementia (n = 8/19), SLI was reduced in presenile dementia (F = 6.80, p = 0.02), while the concentration of H I A A was decreased in senile dementia (F -- 8.50, p -- 0.008). When AD patients were stratified into three groups according to the severity of the disease (Table 2), the SLI levels decreased with the progression of the disease, however, the differences were not statistically significant. A few significant intercorrelations between CSF measures were found; AChE correlated significantly with SLI in AD patients (r = 0.48, p = 0.01), in PD patients (r = 0.66, p = 0.0001) and in controls (r = 0.66, p = 0.001). Similarly, H V A and H I A A were intercorrelated for AD patients (r = 0.56, p -- 0.002), for PD patients (r = 0.75, p = 0.0001) and for controls (Hartikainen et al., 1991). As to clinical correlations, concentrations of HVA were related to Webster's score (r = - 0.46, p -- 0.015) and Hoehn and Yahr score (r = 0.43, p = 0.03) in AD, but not in PD. -
58
P. Hartikainen et al.
c~ "o 9~ t
0 o~
0
0
~
~'~
~
~
oo
0
c~
c~
~
0
~
,~
0
0 ~ ~."
~o~
o ,-~
"o
0
r
.~
+l
-H
-H
~
-H
-H
9
"
09
0
Z
__ Journal o f Neural Transmission 9 Springer-Verlag 1992 Printed in Austria
Neurochemical markers in the cerebrospinal fluid of patients with Alzheimer's disease, Parkinson's disease and amyotrophic lateral sclerosis and normal controls P. Hartikainen 1, K. J. Reinikainen 1, H. Soininen 1, J. Sirvifi2, R. Soikkeli 1, and P. J. Riekkinen 1
1Department of Neurology, and 2Neurochemical Research Laboratory, University of Kuopio, Finland Accepted May 27, 1991
Summary. Several neurotransmitter markers were investigated in the cerebrospinal fluid (CSF) from patients with Alzheimer's disease (AD) (n= 27), Parkinson's disease (PD) (n = 35) and ALS (n=26) and from control subjects (n-- 34) to compare the possible alterations in the biochemical profiles of these different neurodegenerative diseases. The main proportion of the patients represented an early phase of the illness at the time of the diagnosis. Correlations of the degree of dementia and the stage of the disease with CSF measures were evaluated. The CSF levels of somatostatin like-immunoreactivity (SLI) were significantly reduced in AD patients when compared with those of normals and ALS patients. The CSF concentrations of homovanillic acid (HVA) were significantly decreased for PD patients and the decrease focused on the nondemented patients. A trend of decreasing HVA values towards the most advanced stage of Parkinson's disease assessed by Webster's scale was also displayed. The content of 3-methoxy-4-hydroxyphenylglycol (MHPG) in the CSF was higher for ALS patients than for other groups. The lowest 5-hydroxyindoleacetic acid (5HIAA) levels were observed in the PD group and the lowest acetylcholinesterase (ACHE) activities were found in the PD patients with the most severe disease. Changes in CSF measures were too subtle to be beneficial for diagnostic purposes, but adequate for reflecting the different neurochemical profiles of these three degenerative neurological disorders.
Keywords: Cerebrospinal fluid, neurotransmitters, Alzheimer's disease, Parkinson's disease, ALS.
54
P. Hartikainen et al. Introduction
Alzheimer's disease (AD), Parkinson's disease (PD), and amyotrophic lateral sclerosis (ALS) are progressive neurological degenerative disorders of humans in middle and late life. Typically, clinical signs and symptoms are different, although the diseases share some similar features; patients with AD show dementia and later often extrapyramidal signs, PD patients develop a motor disorder with or without dementia and patients with ALS show devasting and weakness of musculature, while dementia in this disease is rare. Neurochemical studies of these diseases have indicated the complexity of the neurotransmitter systems involved. Post mortem studies have consistently shown the deficit of the ascending cholinergic system in AD and the disruption of the nigrostriatal pathways in PD (Rossor et al., 1982; Perry, 1986; Ruberg et al., 1986). The involvement of the ascending cholinergic innervation has also been found in demented PD patients (Dubois et al., 1983). Disturbances of noradrenergic and serotonergic systems have been described both in AD and PD (Rossor and. Mountjoy, 1986; Palmer et al., 1987). Data concerning neurochemical alterations in ALS are limited and difficult to interpret due to widely distributed pathological changes both in the spinal cord and the brain. A proportion of the changes in these neurotransmitter markers reflects to the cerebrospinal fluid (CSF). The most unequivocal finding is a low concentration of homovanillic acid (HVA), a metabolite of dopamine, in the CSF of PD patients (Johansson and Roos, 1967). Many studies have detected reduced cholinesterase (ChE) levels in the CSF from AD and demented PD patients, but not consistently (see Discussion). Therefore, the value of neurotransmitter measurements of the CSF as a diagnostic tool is doubtful. The present study attempts to clarify the profile of several transmitter related parameters in the CSF from patients with AD, PD, and ALS and from controls in relation to the degree of dementia, the stage of the disease and some clinical manifestations. The patients are carefully investigated using the standardized protocol of clinical, neuropsychological, neurophysiological and CSF examinations. Patients and methods Patients
The study included 128 subjects who were admitted to the Neurological Department of Kuopio University Central Hospital for diagnostic or therapeutic evaluation. The subject or, when needed, her/his nearest relative gave informed consent for participation in the study. The study was approved by the local ethics committee. All subjects passed through the following study battery: general physical and neurological examination, assessment of Hachinski Ishaemic score (Rosen et al., 1980), a measurement of blood pressure, ECG, chest X-ray, EEG, spectral analysis of EEG, somatosensoric evoked potentials, visual evoked potentials, brain CT-scan and neuropsychological tests. Laboratory tests consisted of blood cell count, erythrocyte sedimentation rate, blood glucose, serum urea, electrolytes,calcium, phosphate, vitamin B12, erythrocyte folate, liver
Neurochemical markers in the cerebrospinal fluid
55
and thyroid function tests, cholesterol, triglycerides, immunoglobulin (IgG), albumin (alb), lues serology and urine analysis. The routine CSF analysis comprised count of erythrocytes and leukocytes, determination of the total protein content, protein electrophoresis, IgG, alb, IgG/alb ratio, IgG index = (CSF IgG/alb)/(serum IgG/alb). The control population comprised 34 subjects (20 females, 14 males; mean age 49.3 + 16.7 years, range 20-80 years). The subjects had no history of dementia, chronic psychiatric disease or depression and did not use any psychotropic drugs. The subjects were selected from patients investigated due to minor neurological symptoms. In all cases the above mentioned tests were normal. The Alzheimer group consisted of 33 patients (24 females, 9 males; mean age 72.1 + 7.5 years, range 55-85 years) who fulfilled the diagnostic criteria of probable AD according to the definition of the NINCDS-ADRDA Work group (McKhann et al., 1984). All but two patients who were presently diagnosed were at diagnostic investigations. Metabolic, endocrine, nutritional and infectious causes of dementia were excluded. The patients with evidence of cerebrovascular disease according to the modified ischemic scale of Hachinski were excluded. For comparison of CSF findings, we divided the patients into presenile (age < 65 years) and senile dementia groups (age >~ 65 years) according to the onset of the disease and into three groups of clinical severity assessed by the Clinical Dementia Rating scale (CDR) (Hughes et al., 1982) mild (score 0.5-1), moderate (score 2) and severe (score 3) dementia. Extrapyramidal signs and symptoms were evaluated by Webster's scale (Webster, 1968). Five of the patients with severe dementia were receiving neuroleptics, four moderately or severely demented patients and one mildly demented patient were on codergocrin mesylate regimen. One patient with severe dementia had benzodiazepine medication. The duration of the disease from the first signs of the disease was less than 1 year in 2 patients, 1-2 years in 8 patients, 2-5 years in 21 patients, and over 5 years in two patients. Idiopathic PD patients were 12 females and 23 males with mean age 63.9 + 10.4, range 40-80 years. Twenty-eight patients were in diagnostic examinations, 7 patients represented earlier diagnosed and more advanced disease. The stage of the illness was evaluated according to the criteria of Hoehn and Yahr (1967) and clinical signs and symptoms were assessed by Webster's scale (Webster, 1968). The duration of PD was up to 1 year in 9 patients, 1-2 years in 12 patients, 2-5 years in 4 patients, and over 5 years in 7 patients. The severity of dementia was assessed by the CDR scale as in the AD group. Of 10 demented cases (28.6% from all PD patients) 7 patients showed mild and 3 moderate dementia. Eleven patients were on continuous antiparkinsonian treatment; 7 patients received L-dopa, 4 deprenyl, 1 bromocriptine, 1 amantadine, 3 neuroleptics, and 1 patient was on benzodiazepine regimen. The regimen combinations were variable. The patients with L-dopa represented stages IV-V of Hoehn and Yahr scale and most of the other medicated patients were of milder stages of the disease. Twenty-six patients with ALS (8 females, 18 males; mean age 61.9 + 11.2 years, range 28-76 years) were selected for this study. The diagnosis was based on history and clinical examination and it was confirmed b y electromyography showing typical denervation changes, fibrillations, fasciculations and loss of motor units in the peripheral musculature. Furthermore, muscle biopsy and serum creatine kinase values were investigated. Seven patients had predominantly bulbar symptoms and signs, 8 patients represented mainly the spinal form and 11 patients the mixed type of the disease. All ALS patients were at the early stage of the disease in the diagnostic investigations. ALS patients were estimated by the CDR scale: Five of the ALS patients scored 0.5 (questionable dementia), five scored 1 (mild dementia), and all other presented the score 0 (no dementia). In addition, cognitive abilities of ALS patients were assessed by using the Wechsler Adult Intelligence Scale (WAIS) performance IQ scores (PIO). The PIQ score of clinically possible ALS-dementia
56
P. Hartikainen et al.
cases (75.6 + 14.2) differed significantly (p < 0.001) from the scores of non-demented patients (98.6 • 17.9) and age-matched controls (100.8 + 9.1).
CSF samples The subjects were kept on low monoamine diet before the lumbar puncture. The CSF samples were collected after an overnight bed rest and fasting at 08.00-10.00 a.m. the subjects lying in a lateral decubitus position. The medicated patients were drugfree for at least 4 days before the lumbar puncture. The first specimen of CSF, 4 ml, was collected in 1 ml fractions for the routine analysis and the next specimens, together 15 ml in 5 ml fractions were collected for other determinations. Specimens were immediately frozen in liquid nitrogen and stored at - 80 ~ until assayed. Levels of neuropeptides, somatostatin (SLI), adrenocorticotropic hormone (ACTH) and f3-endorphin (BLI), were determined from the first fraction. The activity of acetylcholinesterase (ACHE), concentrations of monoamine metabolites; homovanillic acid (HVA), 5-hydroxyindoleacetic acid (5HIAA), and 3-methoxy-4-hydroxyphenylglycol(MHPG), and protein were determined from the second fraction. The activity of another noradrenergic marker, dopamine-~-hydroxylase (DBH), was measured from the third fraction.
Chemical methods The total protein content of the CSF was measured by the method of Lowry et al. (1951). The levels of monoamine metabolites were determined by using high performance liquid chromatography (HPLC) with electrochemical detection as described previously (Jolkkonen et al., 1987 a). The activity of DBH in the CSF was detected by a radioenzymatic method as reported by Soininen et al. (1984 a). The activity of AChE was determined as described previously by Sirvi6 et al. (1989a). Levels of SLI, BLI, and ACTH immunoreactivity in CSF were measured by radioimmunological assays (Jolkkonen et al., 1987 b, 1990).
Statistical methods The normal distribution of the analysed variables was tested with Kolmogorov-Smirnov test. Differences between groups were evaluated by analysis of variance (ANOVA) adjusted for confounding factors (ANCOVA), or by the non-parametric Kruskall-Wallis test when the assumptions for ANOVA, particularly the equivalency of variance or the normality of the distribution did not fulfil. The post hoc analyses were consequently Student's t-test with Bonferroni correction when appropriate or a modified procedure developed for Kruskall-Wallis analysis (Siegel and Castellan, 1988). In addition, when applying KruskallWallis analysis, the influence of confounding factors was checked by calculating Pearson correlation coefficients for CSF measures. The relationships between clinical measures (Webster's scale and Hoehn & Yahr scale) and CSF variables were assessed by using the non-parametric Spearman correlation test.
Results The m e a n CSF levels o f n e u r o t r a n s m i t t e r related variables o f patients with A D , P D a n d A L S a n d controls are shown in Table 1. Because age, height, weight, a n d season were f o u n d to interfere with the levels C S F measures in our previous w o r k with controls ( H a r t i k a i n e n et al., 1991), all C S F variables but M H P G , D B H , a n d BLI were adjusted for age and, in addition, H V A for height a n d season, A C T H for season a n d weight a n d SLI for weight. The c o n c e n t r a t i o n o f M H P G in the C S F correlated positively with age in n o r m a l subjects (Har-
Neurochemical markers in the cerebrospinal fluid
57
tikainen et al., 1991) and in A D patients (r -- 0.40, p = 0.04), but not in PD (r --- 0.09, p = 0.61) or ALS patients (r = 0.29, p --- 0.10). Seasonal alterations in the CSF levels of DBH and BLI found for normals (Hartikainen et al., 1991) were not observed for any of the disease groups. A C T H values for AD patients (p < 0.05) showed similar seasonal change as did those of controls; the levels were higher in the light than in the dark period (Hartikainen et al., 1991). The total protein content in the CSF showed great interindividual variability in each group. PD patients and ALS patients differed significantly from controls (Table 1). The activity of AChE in the CSF was lowest for A D patients, but the differences did not reach statistical significance between the groups (Table 1). In PD, the activity of AChE in the CSF was reduced in patients with the most advanced disease (Table 4). The AChE activities of demented PD patients were lower than those of non-demented patients (Table 3). The ALS patients had the highest M H P G concentrations of the CSF, and the ALS group differed significantly from the other groups (Table 1). The M H P G levels for demented and non-demented ALS patients were comparable as were all other CSF measures but SLI (Table 5) which, however, did not differ from those of controls (Table 1). Neither did the type of ALS contribute to any changes in neurotransmitter markers of the CSF (data not shown). The DBH activities of the CSF did not differ across the groups (Table 1). As expected, the CSF H V A levels were lowest for PD patients and differed significantly from all other groups (Table 1). In line with previous observations, height, as a covariate in the analysis of variance (ANOVA), showed a significant effect on H V A values of the CSF (F = 19.6, p = 0.0001). The concentrations of 5HIAA were lowest for the PD patients and differed significantly from those of AD patients, but not from the controls (Table 1). SLI values of the CSF were lowest for the AD group and differed significantly from the controls (Table 1). The BLI levels of the CSF did not differ across the groups (Table 1). Neither did the A C T H values of the CSF show any differences between groups (Table 1). In comparison of patients with presenile or senile dementia (n = 8/19), SLI was reduced in presenile dementia (F = 6.80, p = 0.02), while the concentration of H I A A was decreased in senile dementia (F -- 8.50, p -- 0.008). When AD patients were stratified into three groups according to the severity of the disease (Table 2), the SLI levels decreased with the progression of the disease, however, the differences were not statistically significant. A few significant intercorrelations between CSF measures were found; AChE correlated significantly with SLI in AD patients (r = 0.48, p = 0.01), in PD patients (r = 0.66, p = 0.0001) and in controls (r = 0.66, p = 0.001). Similarly, H V A and H I A A were intercorrelated for AD patients (r = 0.56, p -- 0.002), for PD patients (r = 0.75, p = 0.0001) and for controls (Hartikainen et al., 1991). As to clinical correlations, concentrations of HVA were related to Webster's score (r = - 0.46, p -- 0.015) and Hoehn and Yahr score (r = 0.43, p = 0.03) in AD, but not in PD. -
58
P. Hartikainen et al.
c~ "o 9~ t
0 o~
0
0
~
~'~
~
~
oo
0
c~
c~
~
0
~
,~
0
0 ~ ~."
~o~
o ,-~
"o
0
r
.~
+l
-H
-H
~
-H
-H
9
"
09
0
Z
