Eur. Neurol. 18: 140-144 (1979)
Variation of Immunoglobulin G and Total Protein Concentrations during Lumbar Cerebrospinal Fluid Collection G.O. Fossan' and J.L. Larsen Department of Neurology and Department of Roentgenology, Haukeland Sykehus, and Broegelmann Research Laboratory for Microbiology, University of Bergen, Bergen
Key Words. CSF • IgG • Total protein • Variation of concentrations Abstract. In successive lumbar samples of cerebrospinal fluid (CSF), each consisting of 5 ml, immunoglobulin G (IgG) and total protein were determined. Variation was studied by comparing the values in the initial sample with values in succeeding samples. In 13 out of 17 patients, IgG was reduced in sample 2. There was no reduction in the following samples. The concentration of IgG in sample 2 was below 50% of the concentration in sample 1 in 2 patients and below 70% in 9 patients. IgG was also reduced if the concentration in each sample was given as percent of the total protein concentration. The total protein concentration was not significantly reduced during collection of 25 ml of CSF. IgG was increased in sample 2 in 3 out of the 4 patients with neurological diseases since childhood. In 1 patient there was a considerable rise in the total protein concentration after collection of the first 5 ml, without a corresponding rise of IgG concentration. In clinical practice, the portion to which the CSF sample belongs should be noted, and comparison of values from various patients should be made between equivalent samples.
In earlier investigations, cerebrospinal fluid (CSF) immunoglobulin G (IgG) was found to be slightly reduced in some categories of pa tients with neurological diseases (2). In many instances, relatively great volumes of CSF had been collected from these patients. This led to the idea that the IgG concentration might vary during CSF collection. An examination of suc ceeding samples of CSF collected was therefore performed. 1 Fellow of the Norwegian Research Council for Science and the Humanities.
Materials and M ethods
Patients The 17 patients examined, 13 men and 4 women, 18-64 years old, mean age 36 years, underwent pneu moencephalography for various reasons. The dominat ing symptoms were epileptic seizures, intellectual insufficiency, headache and parkinsonism. A head trauma was regarded as the cause of the symptoms in 7 of the patients, a neonatal or prenatal brain lesion was regarded as the cause in 4. Most of the patients had signs of cerebral lesion on clinical examination. Dilated cerebral ventricles were demonstrated in the majority, and some patients showed electroencephalographic abnormalities. 2 patients had symptoms or signs indicating inflammatory disease: 1 multiple sclerosis and 1 myositis. 1 patient (G.A.) will be
Downloaded by: King's College London 137.73.144.138 - 7/25/2018 6:21:34 AM
Introduction
141
Protein Variations during CSF Collection
discussed separately. He had headache as the dominat ing symptom. A left-sided homonymous hemianopia was found on clinical examination, and on pneumo encephalography, no occipital horn could be demon strated at the right cerebral ventricle. No indication of intracranial tumour was found on computer-assisted tomography and cerebral angiography, and no other anomaly. The diagnosis of this patient’s disease is still obscure.
the unstained gamma region into 8 pieces and deter mining the antibody activity in each gel piece after elution with 0.5 ml of phosphate-buffered saline, pH 7.2.
CSF With the patient in the sitting position, the lumbar puncture needle was inserted between the fourth and the fifth lumbar, or between the fifth lumbar and the first sacral, vertebra. Initially, 5 ml of oxygen was injected by means of a syringe and then 5 ml of CSF was withdrawn. Thereafter, 7.5 ml of oxygen was injected and 5 ml of CSF was withdrawn. This opera tion was repeated until at least 5 CSF samples had been collected. The patient was kept sitting during the procedure, the time necessary was between 15 and 30 min. The samples were consecutively numbered from 1 to 5, and higher when more samples were collected. Cells were counted immediately in sample 1. All samples were then stored at - 20 °C. Prior to agarose electrophoresis, the CSF was concentrated using collodium bags, according to Mies (8).
Statistics The Wilcoxon test for paired samples was used for the statistical evaluation.
were
performed
according
to
IgG IgG was determined as described previously (2) using single radial diffusion. Antisera specific for IgG were purchased from Dako-Immunoglobulins, Copen hagen. The lower level of determinations was 0.5 mg/ 100 ml. Determinations were made twice and simul taneously in all samples from each patient. Agarose Electrophoresis Agarose electrophoresis was performed for 1 h as described by Johansson (5) on CSF concentrated 30-40 times. Natural IgG Antibodies to Rabbit Erythrocytes These antibodies were determined by agglutination of rabbit erythrocytes using an antiglobulin test de scribed in detail elsewhere (3). The determination in the agarose after electrophoresis was done by cutting
Results
The distribution of the patients according to IgG, total protein concentrations and number of leucocytes per cubic millimetre in sample 1 is shown in table I. The mean IgG concentra tion was 2.39 mg/100 ml, the mean total pro tein concentration was 42.7 mg/100 ml. The mean IgG concentration, calculated as percent of total protein concentration in each patient, was 5.9%. The titre of antibodies to rabbit erythrocytes varied between 1 and 4. Variation o f IgG in Successive Samples o f CSF Tire distribution of patients according to the IgG concentrations in samples 2 - 5 , expressed as percent of the concentration in sample 1, is presented in figure 1. The IgG concentration in sample 2 was reduced (0.05 > p >0.02) as compared with the concentration in sample 1. None of the other samples differed from sample 1 (p >0.05). In 13 out of the 17 patients, the IgG concentration was lower in sample 2 than in sample 1. The IgG concentration in sample 2 was less than 70% of the concentration in sample 1 in 9 patients and less than 50% in 2 patients. In 4 patients, there was a higher con centration of IgG in sample 2 than in sample 1.
Downloaded by: King's College London 137.73.144.138 - 7/25/2018 6:21:34 AM
Total Protein Determinations Lowry et al. (7).
Cells The number of leucocytes per cubic millimetre of CSF was determined by means of the Fuchs Rosenthal counting chamber.
142
Fossan/ Larsen
Table I. Sample 1 : distribution of patients (n) according to IgG, total protein concentrations and number of leucocytes per cubic millimetre
IgG mg/100 ml
n
Total protein mg/100 ml
n
Leucocytes/mm3
na
1-1.9 2-2.9 3-3.9 4 -4.9 5 -5.9 6-6.9
8 4 3 1 0 1
20-29 30-39 40-49 50-59 60-69
3 6 2 4 2
0 -3 4 -5 6 -7 8-15
9 4 2 1
a Not counted in 1 patient.
Variation o f Total Protein Concentration in Successive Samples o f CSF The distribution of the patients according to the total protein concentration in samples 2—5, expressed as percent of the concentration in sample 1, is presented in figure 2. There was no reduction of total protein concentration in sample 2 similar to that found for IgG in that sample. None of the samples differed (p >0.05) from sample 1, but the number of fluids with a total protein concentration below that seen in
the initial sample was slightly higher than in the number of fluids with total protein concentra tion above that in the initial sample. 1 patient (G.A.) differed clearly from the others; the total protein concentration in sample 1 was within the normal range but there was a consid erable increase in the following samples, with out a corresponding rise in the IgG concentra tion (fig. 3). Discussion
The present results show that the IgG con centration in the CSF may decrease considera bly during the collection of 10 ml of CSF with out a corresponding decrease of the total protein concentration. This was demonstrated by the following results: a greater number of samples had reduced IgG concentration than the number with reduced total protein concen tration; a greater degree of reduction of the concentration of IgG than of total protein was found; the IgG concentration expressed as percent of the total protein concentration in each patient was reduced. Reduced detection of IgG as a consequence of enzymatic degradation is probably not the reason why the IgG concentration was reduced,
Downloaded by: King's College London 137.73.144.138 - 7/25/2018 6:21:34 AM
When IgG in each sample was expressed as percent of total protein, there was also a re duced amount (0.05 > p >0.02) in sample 2 as compared with sample No. 1. Agarose electrophoresis was performed with various samples from a few patients. All electro phoreses showed a normal appearance of the gamma region. There was no variation of the titre of anti bodies to rabbit erythrocytes in successive CSF samples. Determination of antibody activity in small fractions of the gamma region after elec trophoreses of 200 times concentrated and pooled CSF showed that the titres corre sponded with the IgG concentration deter mined by means of radial immunodiffusion.
143
Fig. 1. Distribution of number of patients ac cording to IgG concentration in samples 2 -5 , ex pressed as percent of the concentration in sample 1. The broken line indicates 100%, i.e. the concen tration in sample 1.
Fig. 2. Distribution of number of patients accord ing to total protein concentration in samples 2 -5 , expressed as percent of the concentration in sample 1. The broken line indicates 100%, i.e. the concentration in sample 1.
since the electrophoretic mobility of the IgG was normal and antibody activity was present. A reduction of the antibody activity could not be expected since the two-fold dilution steps are too great for that. These antibodies are represented in the whole gamma region of the electropherogram. As shown earlier (4), they belong to all four subclasses of IgG. Therefore, a reduced IgG in the second sample collected seems the more reasonable explanation of the results. The composition of the CSF may thus be different in the two samples. The source of the two initial samples collect ed is the lower part of tire spinal subarachnoid sac, but the more exact locus for each sample is unknown. If it is assumed that the first sample mainly originates at the close proximity of the puncture needle, the second one partly from tlie spinal subarachnoid sac caudal to the
Fig. 3. Variation of IgG and of total protein con centrations in succeeding CSF samples collected by means of lumbar puncture in patient G.A. The concen trations arc expressed as percent of the concentrations in sample 1.
Downloaded by: King's College London 137.73.144.138 - 7/25/2018 6:21:34 AM
Protein Variations during CSF Collection
144
parts, while the IgG concentrations are similar, in caudal and rostral regions. Regrettably, further diagnostic investigations did not permit the diagnosis of this patient’s disease. The variations of IgG in different samples from some of the patients are of such a magni tude that they might be misleading. In clinical practice, it should therefore be indicated on a CSF sample whether it belongs to the first 5 ml, tiie second 5, or to one of the following por tions collected, and comparison should be made with samples from analogous portions. References
1 Bickerstaff, E.R.: Changes in the cerebrospinal fluid during pneumo-encephalography. Lancet ii: 683-685 (1950). 2 Fossan, G.O. and Tonder, O.: Natural antibodies in human cerebrospinal fluid to rabbit erythrocytes. Int. Archs Allergy appL Immun. 49: 453-463 (1975). 3 Fossan, G.O.: Reduced CSF IgG in patients treated with phenytoin (diphenylhydantoin). Eur. Neurol. ¡4: 426-432 (1976). 4 Fossan, G.O.: The transfer of IgG from serum to CSF, evaluated by means of a naturally occurring antibody. Eur. Neurol. 15: 231-236 (1977). 5 Johansson, B.G.: Agarose gel electrophoresis. Scand. J. clin. Lab. Invest. 29: suppl. 124, pp. 7 19 (1972). 6 Levine, M.C.: Changes in the CSF during pneumo encephalography. Neuroradiology 5: 1 -6 (1973). 7 Lowry, O.H.; Rosebrough, N.J.; Farr, A.L., and Randall, R.J.: Protein measurement with the Folin phenol reagent. J. biol. Chem. 193: 265-275 (1951). 8 Mies, H.-J.: Einengung von Liquor cerebrospinalis als Vorbercitung zur Papierelektrophorese. Klin. Wschr. 31: 159-161 (1953).
Received: April 24, 1978 Accepted: May 9,1978 Dr. G.O. Fossan, Department of Neurology, Haukeland Sykehus, N-5016 Bergen (Norway)
Downloaded by: King's College London 137.73.144.138 - 7/25/2018 6:21:34 AM
puncture needle and the following samples from the more rostral parts of the spinal sub arachnoid space, then the results can be more easily interpreted. They then indicate that the IgG concentration in the lowest part of the spinal subarachnoid sac is lower titan the IgG concentration in the more rostral parts of the spinal subarachnoid sac. An elimination of IgG at a higher rate than the elimination of other proteins in the lower part of tite spinal subarachnoid sac would lead to the results found. So would a reduced entrance of IgG into this region. The role of the local irritant effect of the spinal puncture pro cedure or of the oxygen injected is not clear, but an explanation of tite present results by these factors does not seem very likely. The time span between the first and the second sample was usually not more than 1 min. The increase of IgG in the second and sub sequent samples in 3 out of 4 patients with neurologic disease since childhood indicates an altered composition of the CSF within the spinal canal in these patients. In successive CSF samples, a slight decrease of total protein concentrations has been found (1, 6), reflecting the higher protein concentra tions in the lumbar subarachnoid sac titan in tite cisterna magna. When greater volumes are collected, the fluid will partly come from the cisterna magna. That tendency can be seen in tite present results too. To the authors’ knowl edge, determinations of IgG by means of immu nologic techniques have not been done previ ously in aliquots as small as 5 ml. The finding, in 1 patient, of increasing total protein concentrations without an increase of the IgG concentrations, in successive CSF samples, was unexpected. It probably means that one or more proteins are present in greater concentrations in the more rostral parts of the spinal subarachnoid space than in the caudal
Fossan/Larsen
Variation of Immunoglobulin G and Total Protein Concentrations during Lumbar Cerebrospinal Fluid Collection G.O. Fossan' and J.L. Larsen Department of Neurology and Department of Roentgenology, Haukeland Sykehus, and Broegelmann Research Laboratory for Microbiology, University of Bergen, Bergen
Key Words. CSF • IgG • Total protein • Variation of concentrations Abstract. In successive lumbar samples of cerebrospinal fluid (CSF), each consisting of 5 ml, immunoglobulin G (IgG) and total protein were determined. Variation was studied by comparing the values in the initial sample with values in succeeding samples. In 13 out of 17 patients, IgG was reduced in sample 2. There was no reduction in the following samples. The concentration of IgG in sample 2 was below 50% of the concentration in sample 1 in 2 patients and below 70% in 9 patients. IgG was also reduced if the concentration in each sample was given as percent of the total protein concentration. The total protein concentration was not significantly reduced during collection of 25 ml of CSF. IgG was increased in sample 2 in 3 out of the 4 patients with neurological diseases since childhood. In 1 patient there was a considerable rise in the total protein concentration after collection of the first 5 ml, without a corresponding rise of IgG concentration. In clinical practice, the portion to which the CSF sample belongs should be noted, and comparison of values from various patients should be made between equivalent samples.
In earlier investigations, cerebrospinal fluid (CSF) immunoglobulin G (IgG) was found to be slightly reduced in some categories of pa tients with neurological diseases (2). In many instances, relatively great volumes of CSF had been collected from these patients. This led to the idea that the IgG concentration might vary during CSF collection. An examination of suc ceeding samples of CSF collected was therefore performed. 1 Fellow of the Norwegian Research Council for Science and the Humanities.
Materials and M ethods
Patients The 17 patients examined, 13 men and 4 women, 18-64 years old, mean age 36 years, underwent pneu moencephalography for various reasons. The dominat ing symptoms were epileptic seizures, intellectual insufficiency, headache and parkinsonism. A head trauma was regarded as the cause of the symptoms in 7 of the patients, a neonatal or prenatal brain lesion was regarded as the cause in 4. Most of the patients had signs of cerebral lesion on clinical examination. Dilated cerebral ventricles were demonstrated in the majority, and some patients showed electroencephalographic abnormalities. 2 patients had symptoms or signs indicating inflammatory disease: 1 multiple sclerosis and 1 myositis. 1 patient (G.A.) will be
Downloaded by: King's College London 137.73.144.138 - 7/25/2018 6:21:34 AM
Introduction
141
Protein Variations during CSF Collection
discussed separately. He had headache as the dominat ing symptom. A left-sided homonymous hemianopia was found on clinical examination, and on pneumo encephalography, no occipital horn could be demon strated at the right cerebral ventricle. No indication of intracranial tumour was found on computer-assisted tomography and cerebral angiography, and no other anomaly. The diagnosis of this patient’s disease is still obscure.
the unstained gamma region into 8 pieces and deter mining the antibody activity in each gel piece after elution with 0.5 ml of phosphate-buffered saline, pH 7.2.
CSF With the patient in the sitting position, the lumbar puncture needle was inserted between the fourth and the fifth lumbar, or between the fifth lumbar and the first sacral, vertebra. Initially, 5 ml of oxygen was injected by means of a syringe and then 5 ml of CSF was withdrawn. Thereafter, 7.5 ml of oxygen was injected and 5 ml of CSF was withdrawn. This opera tion was repeated until at least 5 CSF samples had been collected. The patient was kept sitting during the procedure, the time necessary was between 15 and 30 min. The samples were consecutively numbered from 1 to 5, and higher when more samples were collected. Cells were counted immediately in sample 1. All samples were then stored at - 20 °C. Prior to agarose electrophoresis, the CSF was concentrated using collodium bags, according to Mies (8).
Statistics The Wilcoxon test for paired samples was used for the statistical evaluation.
were
performed
according
to
IgG IgG was determined as described previously (2) using single radial diffusion. Antisera specific for IgG were purchased from Dako-Immunoglobulins, Copen hagen. The lower level of determinations was 0.5 mg/ 100 ml. Determinations were made twice and simul taneously in all samples from each patient. Agarose Electrophoresis Agarose electrophoresis was performed for 1 h as described by Johansson (5) on CSF concentrated 30-40 times. Natural IgG Antibodies to Rabbit Erythrocytes These antibodies were determined by agglutination of rabbit erythrocytes using an antiglobulin test de scribed in detail elsewhere (3). The determination in the agarose after electrophoresis was done by cutting
Results
The distribution of the patients according to IgG, total protein concentrations and number of leucocytes per cubic millimetre in sample 1 is shown in table I. The mean IgG concentra tion was 2.39 mg/100 ml, the mean total pro tein concentration was 42.7 mg/100 ml. The mean IgG concentration, calculated as percent of total protein concentration in each patient, was 5.9%. The titre of antibodies to rabbit erythrocytes varied between 1 and 4. Variation o f IgG in Successive Samples o f CSF Tire distribution of patients according to the IgG concentrations in samples 2 - 5 , expressed as percent of the concentration in sample 1, is presented in figure 1. The IgG concentration in sample 2 was reduced (0.05 > p >0.02) as compared with the concentration in sample 1. None of the other samples differed from sample 1 (p >0.05). In 13 out of the 17 patients, the IgG concentration was lower in sample 2 than in sample 1. The IgG concentration in sample 2 was less than 70% of the concentration in sample 1 in 9 patients and less than 50% in 2 patients. In 4 patients, there was a higher con centration of IgG in sample 2 than in sample 1.
Downloaded by: King's College London 137.73.144.138 - 7/25/2018 6:21:34 AM
Total Protein Determinations Lowry et al. (7).
Cells The number of leucocytes per cubic millimetre of CSF was determined by means of the Fuchs Rosenthal counting chamber.
142
Fossan/ Larsen
Table I. Sample 1 : distribution of patients (n) according to IgG, total protein concentrations and number of leucocytes per cubic millimetre
IgG mg/100 ml
n
Total protein mg/100 ml
n
Leucocytes/mm3
na
1-1.9 2-2.9 3-3.9 4 -4.9 5 -5.9 6-6.9
8 4 3 1 0 1
20-29 30-39 40-49 50-59 60-69
3 6 2 4 2
0 -3 4 -5 6 -7 8-15
9 4 2 1
a Not counted in 1 patient.
Variation o f Total Protein Concentration in Successive Samples o f CSF The distribution of the patients according to the total protein concentration in samples 2—5, expressed as percent of the concentration in sample 1, is presented in figure 2. There was no reduction of total protein concentration in sample 2 similar to that found for IgG in that sample. None of the samples differed (p >0.05) from sample 1, but the number of fluids with a total protein concentration below that seen in
the initial sample was slightly higher than in the number of fluids with total protein concentra tion above that in the initial sample. 1 patient (G.A.) differed clearly from the others; the total protein concentration in sample 1 was within the normal range but there was a consid erable increase in the following samples, with out a corresponding rise in the IgG concentra tion (fig. 3). Discussion
The present results show that the IgG con centration in the CSF may decrease considera bly during the collection of 10 ml of CSF with out a corresponding decrease of the total protein concentration. This was demonstrated by the following results: a greater number of samples had reduced IgG concentration than the number with reduced total protein concen tration; a greater degree of reduction of the concentration of IgG than of total protein was found; the IgG concentration expressed as percent of the total protein concentration in each patient was reduced. Reduced detection of IgG as a consequence of enzymatic degradation is probably not the reason why the IgG concentration was reduced,
Downloaded by: King's College London 137.73.144.138 - 7/25/2018 6:21:34 AM
When IgG in each sample was expressed as percent of total protein, there was also a re duced amount (0.05 > p >0.02) in sample 2 as compared with sample No. 1. Agarose electrophoresis was performed with various samples from a few patients. All electro phoreses showed a normal appearance of the gamma region. There was no variation of the titre of anti bodies to rabbit erythrocytes in successive CSF samples. Determination of antibody activity in small fractions of the gamma region after elec trophoreses of 200 times concentrated and pooled CSF showed that the titres corre sponded with the IgG concentration deter mined by means of radial immunodiffusion.
143
Fig. 1. Distribution of number of patients ac cording to IgG concentration in samples 2 -5 , ex pressed as percent of the concentration in sample 1. The broken line indicates 100%, i.e. the concen tration in sample 1.
Fig. 2. Distribution of number of patients accord ing to total protein concentration in samples 2 -5 , expressed as percent of the concentration in sample 1. The broken line indicates 100%, i.e. the concentration in sample 1.
since the electrophoretic mobility of the IgG was normal and antibody activity was present. A reduction of the antibody activity could not be expected since the two-fold dilution steps are too great for that. These antibodies are represented in the whole gamma region of the electropherogram. As shown earlier (4), they belong to all four subclasses of IgG. Therefore, a reduced IgG in the second sample collected seems the more reasonable explanation of the results. The composition of the CSF may thus be different in the two samples. The source of the two initial samples collect ed is the lower part of tire spinal subarachnoid sac, but the more exact locus for each sample is unknown. If it is assumed that the first sample mainly originates at the close proximity of the puncture needle, the second one partly from tlie spinal subarachnoid sac caudal to the
Fig. 3. Variation of IgG and of total protein con centrations in succeeding CSF samples collected by means of lumbar puncture in patient G.A. The concen trations arc expressed as percent of the concentrations in sample 1.
Downloaded by: King's College London 137.73.144.138 - 7/25/2018 6:21:34 AM
Protein Variations during CSF Collection
144
parts, while the IgG concentrations are similar, in caudal and rostral regions. Regrettably, further diagnostic investigations did not permit the diagnosis of this patient’s disease. The variations of IgG in different samples from some of the patients are of such a magni tude that they might be misleading. In clinical practice, it should therefore be indicated on a CSF sample whether it belongs to the first 5 ml, tiie second 5, or to one of the following por tions collected, and comparison should be made with samples from analogous portions. References
1 Bickerstaff, E.R.: Changes in the cerebrospinal fluid during pneumo-encephalography. Lancet ii: 683-685 (1950). 2 Fossan, G.O. and Tonder, O.: Natural antibodies in human cerebrospinal fluid to rabbit erythrocytes. Int. Archs Allergy appL Immun. 49: 453-463 (1975). 3 Fossan, G.O.: Reduced CSF IgG in patients treated with phenytoin (diphenylhydantoin). Eur. Neurol. ¡4: 426-432 (1976). 4 Fossan, G.O.: The transfer of IgG from serum to CSF, evaluated by means of a naturally occurring antibody. Eur. Neurol. 15: 231-236 (1977). 5 Johansson, B.G.: Agarose gel electrophoresis. Scand. J. clin. Lab. Invest. 29: suppl. 124, pp. 7 19 (1972). 6 Levine, M.C.: Changes in the CSF during pneumo encephalography. Neuroradiology 5: 1 -6 (1973). 7 Lowry, O.H.; Rosebrough, N.J.; Farr, A.L., and Randall, R.J.: Protein measurement with the Folin phenol reagent. J. biol. Chem. 193: 265-275 (1951). 8 Mies, H.-J.: Einengung von Liquor cerebrospinalis als Vorbercitung zur Papierelektrophorese. Klin. Wschr. 31: 159-161 (1953).
Received: April 24, 1978 Accepted: May 9,1978 Dr. G.O. Fossan, Department of Neurology, Haukeland Sykehus, N-5016 Bergen (Norway)
Downloaded by: King's College London 137.73.144.138 - 7/25/2018 6:21:34 AM
puncture needle and the following samples from the more rostral parts of the spinal sub arachnoid space, then the results can be more easily interpreted. They then indicate that the IgG concentration in the lowest part of the spinal subarachnoid sac is lower titan the IgG concentration in the more rostral parts of the spinal subarachnoid sac. An elimination of IgG at a higher rate than the elimination of other proteins in the lower part of tite spinal subarachnoid sac would lead to the results found. So would a reduced entrance of IgG into this region. The role of the local irritant effect of the spinal puncture pro cedure or of the oxygen injected is not clear, but an explanation of tite present results by these factors does not seem very likely. The time span between the first and the second sample was usually not more than 1 min. The increase of IgG in the second and sub sequent samples in 3 out of 4 patients with neurologic disease since childhood indicates an altered composition of the CSF within the spinal canal in these patients. In successive CSF samples, a slight decrease of total protein concentrations has been found (1, 6), reflecting the higher protein concentra tions in the lumbar subarachnoid sac titan in tite cisterna magna. When greater volumes are collected, the fluid will partly come from the cisterna magna. That tendency can be seen in tite present results too. To the authors’ knowl edge, determinations of IgG by means of immu nologic techniques have not been done previ ously in aliquots as small as 5 ml. The finding, in 1 patient, of increasing total protein concentrations without an increase of the IgG concentrations, in successive CSF samples, was unexpected. It probably means that one or more proteins are present in greater concentrations in the more rostral parts of the spinal subarachnoid space than in the caudal
Fossan/Larsen
