.
Cor relation of magn.etic resonance imagi and CSF findingsh patients with acute monosymptomatic optic neuritis ~
Frederiksen JL, Larsson HBW, Olesen J. Correlation of magnetic resonance imaging and CSF findings in patients with acute monosymptomatic optic neuritis. Acta Neurol Scand 1992: 86: 317-322.
I
Acute monosymptomatic optic neuritis (AMON) may be the first indication of multiple sclerosis (MS), and this sign offers a special opportunity to study the very early clinical stages of MS. This prospective investigation compares results of C S F findings and magnetic resonance imaging (MRI) in a large, homogeneous and well-defined group of patients with AMON. Of 68 consecutively referred patients, 11 had clinically definite MS, another 5 refused a lumbar puncture, and 7 could not participate for various reasons. With the remaining 45 untreated patients, aged 12-52 (mean 31) years, with idiopathic AMON, we have studied interrelationships of C S F findings (leucocyte count, IgG-index and in 29 of the patients oligoclonal bands (OB)) and MRI. Lumbar puncture and MRI were performed within median 24 and 16 days of onset, respectively. In the CSF one or more abnormalities (in 17/45 = 38% pleocytosis, in 16/45 = 36% increased IgG-index, and in 20129 = 69% OB) was found in a total of 23/29 = 79% of patients. MRI at 1.5 T (double SE and IR sequences) showed multiple cerebral lesions in 65% of patients. A significant relation was observed between results of MRI and leucocyte count (p < 0.05) and between results of MRI and IgG-index (p < 0.05), but not between results of MRI and OB (p > 0.20). Over a median observation period of 27 months, 13 patients developed clinically definite MS. All of these patients had lesions on MRI at onset, illustrating the prognostic importance of MRI findings. Results of CSF had until now no marked predictive value for developing clinically definite MS.
Acute monosymptomatic optic neuritis (AMON) may be regarded as a specific disease entity, or ON may be viewed as a form of MS representing one end of a continuum with fully developed clinically definite MS (CDMS) at the other end (1-3). In a previous study using magnetic resonance imaging (MRI) of the brain, we gave evidence of idiopathic AMON being from the very onset regarded as a first manifestation of MS in approximately 2/3 of the 65 cases (4). The purpose of this study was to characterize CSF findings in patients with AMON, to consider them in the context of MRI findings, and to discuss the implications of CSF and MRI abnormalities with AMON in terms of disease pathogenesis and prognosis. Until now there have been no parameters of the CSF specific for MS. Quantitative determination of immunoglobulin G (IgG) in CSF and the IgGindex, that is CSF IgG x serum albumin/serum IgG x CSF albumin, are among the best established
J. L. Frederiksen’, H. B. W. Larsson’, J. Olesen
’
’
Departments of Neurology, Gentofte Hospital, Magnetic Resonance, Hvidovre Hospital, Copenhagen, Denmark
Jette L. Frederiksen, Department of Neurology, Gentofte University Hospital, DK-2900 Hellerup, Denmark Accepted for publication February 5, 1992
laboratory tests in MS and are included as laboratory parameters in Poser’s diagnostic criteria of MS (3). Additional supportative evidence for the diagnosis of MS is the presence of two or more bands of restricted electrophoretic mobility in the gammaglobulin area, i.e. oligoclonal bands (OB) (3). Mononuclear pleocytosis in the CSF is a rather weak and very unspecific indication of demyelinating disease (5). Material and methods
Referral and selection of patients has previously been described in detail (4). In short, ophthalmologists and neurologists from a well-defined catchment area with 1.5 million inhabitants were asked to immediately refer patients with acute symptoms and signs of ON to our neurological department for neurological examination and a lumbar puncture. Based on 317
Frederiksen et al. incidence estimates, most patients seem to have been referred. During nearly 2 years 68 consecutive patients with symptoms lasting less than 14 days fulfilled clinical criteria of ON (4). By careful neurological and ophthalmological examinations and excessive screening with relevant blood tests, other causes of the acute visual symptoms were ruled out. Eleven patients with known etiology of ON, all of whom had CDMS, were not enrolled into this study. Seven patients could not participate in the study for various reasons (4). Five patients refused a lumbar puncture. The remaining 45 patients (30 women, 15 men), aged 12-52 (median 3 1) years, participated in this comparative study of MRI and CSF findings. All patients were untreated at the time of examination. Patients were admitted to our neurological department for one day for lumbar puncture, which was performed with a 22-gauge needle by the same physician (JF) within 8-49 days, median 24 days, from onset of symptoms. After measuring the initial pressure of the CSF, 15-20ml of the CSF was withdrawn and analyzed for content of cells, glucose and total protein as well as IgG-index and OB. Concomitant blood was obtained, too. The method initially used for evaluating OB, the sensitivity of which showed up to be unsatisfactory low, required concentrated CSF. Therefore, CSF was only available for further analysis of OB in 29 of the patients. The investigators were blinded to the results of the MRI studies.
found in 2 patients (neuroborreliosis, cerebral abscess) of 17 patients with other neurological diseases. MRI
The method, the scoring system and the control group (34 healthy volunteers aged 20-55 years) are described in detail in a previous paper (4). The MRI study was carried out within 3-49 (median 16) days from onset of symptoms. MRI was performed at 1.5 T with 2 sequences (slice thickness 4 mm, voxel size 1.2 x 1.2 x 4 mm’): double spin-echo (TR = 1.8 s, T E = 30 and 90 ms, 24 slices axially) and inversion recovery (TR = 2.45 s, TI = 400 ms, TE = 30 ms, 10 slices sagittally). All the MRIs were reviewed blindly by the same investigator (HB WL) who was unaware of the results of the examination of CSF. The MRI-data was rated on a scale from 0 (normal MRI), increasing in steps of 1 to 6, depending on the number and size of lesions. By estimating slice by slice the length, width and depth (slice thickness 4 mm) of plaques, we made a quantitation of the total area of MRI abnormality in the cerebrum for each patient. Statistics
For statistical analysis of dichotomous variables the Fisher’s Exact Test was used. Results
Leucocyte count in CSF
Upper normal limit for leucocyte count was 3 x 106/1 CSF. lgG-index
Albumin and IgG were determined by electroimmunoassay (6). Paired serum and CSF samples, obtained at the same time, were analyzed in the same run. Reference values of IgG-index in normals (0.360.76) were similar to those of Link et al. (7). Oligoclonal bands
OB was detected in CSF by isoelectric focusing in agarose using double-antibody peroxidase labeling and avidin-biotin amplification (8). The detection of OB was done by visual inspection. The CSF was only deemed to be positive for OB, if OB was not present in the serum and if the CSF contained two or more discrete bands in the gammaglobulin region. The sensitivity of this method of detecting OB in CSF in patients with CDMS was 21/23 = 91%. Concerning the specificity of the method, OB was 318
Cerebral lesions were demonstrated by MRI in 28/43 (65%) of patients with AMON. OB was a more sensitive method, i.e. 20/29 (69%) abnormal than IgG-index i.e. 16/45 (36%) in revealing abnormalities in the CSF in patients with AMON. Mononuclear pleocytosis in the CSF was found in 17/45 (38%) of patients. In only 4 patients did the mononuclear leucocyte count exceed lo7 per liter. There was a trend towards the highest values being observed in younger patients. Only 2/15 (13 2) of patients with a normal MRI (i.e. score 0) showed mononuclear pleocytosis in the CSF. A significant relation was observed between results of MRI findings and of leucocyte count in the CSF (p 0.20). Interestingly, IgG-index was normal (i.e. 20.76) in all but 2 patients with a normal MRI (i.e. score 0). Conversely, MRI revealed lesions in 14/28 (50%) patients with a normal IgG-index. Results of MRI and of IgGindex were significantly related (p < 0.05) (Fig. 2). As it appears from Fig. 3, results of MRI and of analysis for OB were not significantly related (p>0.20). Concomitant bands in CSF and serum
ab MRI, CSF & AMON
Leucocyte count (csf ) (rnillions/l) 28
-
IgG-index
I I
normal
I I I I
24 -
normal
r
20
-
16
-
abnormal
MRI
I I
I
IgG-index
I I I
I
3.0
I
I
I
I
I I
II
2.5 -
I I I
2.0 -
I I I
1.5-O 4
r
P
I
0
t-
0
1
2
3
4
5
6
Total MRI score Fig. 1. The distribution of 43 patients with AMON, according to MRI of the brain and leucocyte count in the CSF. In additionally 2 patients one or both tests were not performed. Normal MRI (score 0) is accompanied by normal leucocyte count in CSF in all but 2 patients. Approximately half of the patients with minor abnormalities on MRI (score 2-3) show pleocytosis, whereas the small subgroup of patients with major abnormalities on MRI (score 5-6) shows pleocytosis. The results of the two tests are significantly related (p < 0.05).
were not observed in any patient. In the group of 29 patients with AMON examined for OB, normal results of IgG-index was observed in all but one patient without OB. On the other hand, OB was dem-
present o ligoc lona I bands absent
normal
abnormal
MRI
Fig. 3. The results of MRI of the brain compared to the results of OB in 29 patients with AMON. No significant relation between results of MRI and OB is observed (p > 0.20).
- . 0
1
2
3.
4
5 6 Total MRI score
Fig. 2. The number of patients (in total 44) with acute AMON classified according to the results of MRI of the brain (rated from 0 (i.e. normal) to 6 ) and IgG-index, respectively. One additional patient did not perform both tests. The results of the two tests are significantly related (p < 0.05).
onstrated in 15 patients with normal IgG-index. However, the results of the two tests were not significantly related (p>0.20). A comparison of the combined CSF findings (leucocyte count, IgG-index and OB) with the results of MRI was done in 29 of the patients. No signdicant relation between these results was observed (p>0.20) (Fig. 4). Four patients aged more than 40 years had minor lesions on MRI, but these MRIs did not differ significantly from the MRIs of an age- and sex-matched control group (Table 1). If, however, these 4 patients move from the category “abnormal MRI” to the category “normal MRI”, the relationship between CSF findings and MRI findings will still not be significant. Table2 lists the results of CSF and (when performed) MRI findings obtained in previous studies. In the 28 patients with abnormal MRI, we estimated the total area of MRI abnormality (mean 2325 mm’, SD = 4967 mm’). There was neither significant correlation between the plaque volume and leucocyte count (r = 0.191, t = 0.992, f = 26) nor between the plaque volume and IgG-index (r = - 0.284, t = - 0.158, f = 26). 319
Frederiksen et a].
one or more abnormal
10
I 13
4
2
CSF analyses normal
normal
abnormal
MRI Fig. 4 . The results of MRI of the brain compared to results obtained from a combination of CSF-analyses (leucocyte count, IgG-index and OB) in 29 patients with AMON. The statistical analysis revealed no significant relation between the two parameters (p> 0.20).
Table 1. Four patients aged more than 40 years had lesions on MRI, but these MRls did not differ significantly from the MRls of an age- and sexmatched control group. N=normal, A=abnormal, ND=not done Sex
Age
Leucocyte count
IgG-index
06
MRI
F M
43 52 49 42
A N
A
N ND ND ND
N N N N
M F
N N
N N N
During the observation period from the time of the lumbar puncture and MRI to the recent clinical evaluation of the patients (median 27 months, range 16-38 months), 13 patients developed CDMS. They all belonged to the subgroup of patients with abnormal MRI at onset. The relation between the development of CDMS and the results of MRI is statistically significant (p
Cor relation of magn.etic resonance imagi and CSF findingsh patients with acute monosymptomatic optic neuritis ~
Frederiksen JL, Larsson HBW, Olesen J. Correlation of magnetic resonance imaging and CSF findings in patients with acute monosymptomatic optic neuritis. Acta Neurol Scand 1992: 86: 317-322.
I
Acute monosymptomatic optic neuritis (AMON) may be the first indication of multiple sclerosis (MS), and this sign offers a special opportunity to study the very early clinical stages of MS. This prospective investigation compares results of C S F findings and magnetic resonance imaging (MRI) in a large, homogeneous and well-defined group of patients with AMON. Of 68 consecutively referred patients, 11 had clinically definite MS, another 5 refused a lumbar puncture, and 7 could not participate for various reasons. With the remaining 45 untreated patients, aged 12-52 (mean 31) years, with idiopathic AMON, we have studied interrelationships of C S F findings (leucocyte count, IgG-index and in 29 of the patients oligoclonal bands (OB)) and MRI. Lumbar puncture and MRI were performed within median 24 and 16 days of onset, respectively. In the CSF one or more abnormalities (in 17/45 = 38% pleocytosis, in 16/45 = 36% increased IgG-index, and in 20129 = 69% OB) was found in a total of 23/29 = 79% of patients. MRI at 1.5 T (double SE and IR sequences) showed multiple cerebral lesions in 65% of patients. A significant relation was observed between results of MRI and leucocyte count (p < 0.05) and between results of MRI and IgG-index (p < 0.05), but not between results of MRI and OB (p > 0.20). Over a median observation period of 27 months, 13 patients developed clinically definite MS. All of these patients had lesions on MRI at onset, illustrating the prognostic importance of MRI findings. Results of CSF had until now no marked predictive value for developing clinically definite MS.
Acute monosymptomatic optic neuritis (AMON) may be regarded as a specific disease entity, or ON may be viewed as a form of MS representing one end of a continuum with fully developed clinically definite MS (CDMS) at the other end (1-3). In a previous study using magnetic resonance imaging (MRI) of the brain, we gave evidence of idiopathic AMON being from the very onset regarded as a first manifestation of MS in approximately 2/3 of the 65 cases (4). The purpose of this study was to characterize CSF findings in patients with AMON, to consider them in the context of MRI findings, and to discuss the implications of CSF and MRI abnormalities with AMON in terms of disease pathogenesis and prognosis. Until now there have been no parameters of the CSF specific for MS. Quantitative determination of immunoglobulin G (IgG) in CSF and the IgGindex, that is CSF IgG x serum albumin/serum IgG x CSF albumin, are among the best established
J. L. Frederiksen’, H. B. W. Larsson’, J. Olesen
’
’
Departments of Neurology, Gentofte Hospital, Magnetic Resonance, Hvidovre Hospital, Copenhagen, Denmark
Jette L. Frederiksen, Department of Neurology, Gentofte University Hospital, DK-2900 Hellerup, Denmark Accepted for publication February 5, 1992
laboratory tests in MS and are included as laboratory parameters in Poser’s diagnostic criteria of MS (3). Additional supportative evidence for the diagnosis of MS is the presence of two or more bands of restricted electrophoretic mobility in the gammaglobulin area, i.e. oligoclonal bands (OB) (3). Mononuclear pleocytosis in the CSF is a rather weak and very unspecific indication of demyelinating disease (5). Material and methods
Referral and selection of patients has previously been described in detail (4). In short, ophthalmologists and neurologists from a well-defined catchment area with 1.5 million inhabitants were asked to immediately refer patients with acute symptoms and signs of ON to our neurological department for neurological examination and a lumbar puncture. Based on 317
Frederiksen et al. incidence estimates, most patients seem to have been referred. During nearly 2 years 68 consecutive patients with symptoms lasting less than 14 days fulfilled clinical criteria of ON (4). By careful neurological and ophthalmological examinations and excessive screening with relevant blood tests, other causes of the acute visual symptoms were ruled out. Eleven patients with known etiology of ON, all of whom had CDMS, were not enrolled into this study. Seven patients could not participate in the study for various reasons (4). Five patients refused a lumbar puncture. The remaining 45 patients (30 women, 15 men), aged 12-52 (median 3 1) years, participated in this comparative study of MRI and CSF findings. All patients were untreated at the time of examination. Patients were admitted to our neurological department for one day for lumbar puncture, which was performed with a 22-gauge needle by the same physician (JF) within 8-49 days, median 24 days, from onset of symptoms. After measuring the initial pressure of the CSF, 15-20ml of the CSF was withdrawn and analyzed for content of cells, glucose and total protein as well as IgG-index and OB. Concomitant blood was obtained, too. The method initially used for evaluating OB, the sensitivity of which showed up to be unsatisfactory low, required concentrated CSF. Therefore, CSF was only available for further analysis of OB in 29 of the patients. The investigators were blinded to the results of the MRI studies.
found in 2 patients (neuroborreliosis, cerebral abscess) of 17 patients with other neurological diseases. MRI
The method, the scoring system and the control group (34 healthy volunteers aged 20-55 years) are described in detail in a previous paper (4). The MRI study was carried out within 3-49 (median 16) days from onset of symptoms. MRI was performed at 1.5 T with 2 sequences (slice thickness 4 mm, voxel size 1.2 x 1.2 x 4 mm’): double spin-echo (TR = 1.8 s, T E = 30 and 90 ms, 24 slices axially) and inversion recovery (TR = 2.45 s, TI = 400 ms, TE = 30 ms, 10 slices sagittally). All the MRIs were reviewed blindly by the same investigator (HB WL) who was unaware of the results of the examination of CSF. The MRI-data was rated on a scale from 0 (normal MRI), increasing in steps of 1 to 6, depending on the number and size of lesions. By estimating slice by slice the length, width and depth (slice thickness 4 mm) of plaques, we made a quantitation of the total area of MRI abnormality in the cerebrum for each patient. Statistics
For statistical analysis of dichotomous variables the Fisher’s Exact Test was used. Results
Leucocyte count in CSF
Upper normal limit for leucocyte count was 3 x 106/1 CSF. lgG-index
Albumin and IgG were determined by electroimmunoassay (6). Paired serum and CSF samples, obtained at the same time, were analyzed in the same run. Reference values of IgG-index in normals (0.360.76) were similar to those of Link et al. (7). Oligoclonal bands
OB was detected in CSF by isoelectric focusing in agarose using double-antibody peroxidase labeling and avidin-biotin amplification (8). The detection of OB was done by visual inspection. The CSF was only deemed to be positive for OB, if OB was not present in the serum and if the CSF contained two or more discrete bands in the gammaglobulin region. The sensitivity of this method of detecting OB in CSF in patients with CDMS was 21/23 = 91%. Concerning the specificity of the method, OB was 318
Cerebral lesions were demonstrated by MRI in 28/43 (65%) of patients with AMON. OB was a more sensitive method, i.e. 20/29 (69%) abnormal than IgG-index i.e. 16/45 (36%) in revealing abnormalities in the CSF in patients with AMON. Mononuclear pleocytosis in the CSF was found in 17/45 (38%) of patients. In only 4 patients did the mononuclear leucocyte count exceed lo7 per liter. There was a trend towards the highest values being observed in younger patients. Only 2/15 (13 2) of patients with a normal MRI (i.e. score 0) showed mononuclear pleocytosis in the CSF. A significant relation was observed between results of MRI findings and of leucocyte count in the CSF (p 0.20). Interestingly, IgG-index was normal (i.e. 20.76) in all but 2 patients with a normal MRI (i.e. score 0). Conversely, MRI revealed lesions in 14/28 (50%) patients with a normal IgG-index. Results of MRI and of IgGindex were significantly related (p < 0.05) (Fig. 2). As it appears from Fig. 3, results of MRI and of analysis for OB were not significantly related (p>0.20). Concomitant bands in CSF and serum
ab MRI, CSF & AMON
Leucocyte count (csf ) (rnillions/l) 28
-
IgG-index
I I
normal
I I I I
24 -
normal
r
20
-
16
-
abnormal
MRI
I I
I
IgG-index
I I I
I
3.0
I
I
I
I
I I
II
2.5 -
I I I
2.0 -
I I I
1.5-O 4
r
P
I
0
t-
0
1
2
3
4
5
6
Total MRI score Fig. 1. The distribution of 43 patients with AMON, according to MRI of the brain and leucocyte count in the CSF. In additionally 2 patients one or both tests were not performed. Normal MRI (score 0) is accompanied by normal leucocyte count in CSF in all but 2 patients. Approximately half of the patients with minor abnormalities on MRI (score 2-3) show pleocytosis, whereas the small subgroup of patients with major abnormalities on MRI (score 5-6) shows pleocytosis. The results of the two tests are significantly related (p < 0.05).
were not observed in any patient. In the group of 29 patients with AMON examined for OB, normal results of IgG-index was observed in all but one patient without OB. On the other hand, OB was dem-
present o ligoc lona I bands absent
normal
abnormal
MRI
Fig. 3. The results of MRI of the brain compared to the results of OB in 29 patients with AMON. No significant relation between results of MRI and OB is observed (p > 0.20).
- . 0
1
2
3.
4
5 6 Total MRI score
Fig. 2. The number of patients (in total 44) with acute AMON classified according to the results of MRI of the brain (rated from 0 (i.e. normal) to 6 ) and IgG-index, respectively. One additional patient did not perform both tests. The results of the two tests are significantly related (p < 0.05).
onstrated in 15 patients with normal IgG-index. However, the results of the two tests were not significantly related (p>0.20). A comparison of the combined CSF findings (leucocyte count, IgG-index and OB) with the results of MRI was done in 29 of the patients. No signdicant relation between these results was observed (p>0.20) (Fig. 4). Four patients aged more than 40 years had minor lesions on MRI, but these MRIs did not differ significantly from the MRIs of an age- and sex-matched control group (Table 1). If, however, these 4 patients move from the category “abnormal MRI” to the category “normal MRI”, the relationship between CSF findings and MRI findings will still not be significant. Table2 lists the results of CSF and (when performed) MRI findings obtained in previous studies. In the 28 patients with abnormal MRI, we estimated the total area of MRI abnormality (mean 2325 mm’, SD = 4967 mm’). There was neither significant correlation between the plaque volume and leucocyte count (r = 0.191, t = 0.992, f = 26) nor between the plaque volume and IgG-index (r = - 0.284, t = - 0.158, f = 26). 319
Frederiksen et a].
one or more abnormal
10
I 13
4
2
CSF analyses normal
normal
abnormal
MRI Fig. 4 . The results of MRI of the brain compared to results obtained from a combination of CSF-analyses (leucocyte count, IgG-index and OB) in 29 patients with AMON. The statistical analysis revealed no significant relation between the two parameters (p> 0.20).
Table 1. Four patients aged more than 40 years had lesions on MRI, but these MRls did not differ significantly from the MRls of an age- and sexmatched control group. N=normal, A=abnormal, ND=not done Sex
Age
Leucocyte count
IgG-index
06
MRI
F M
43 52 49 42
A N
A
N ND ND ND
N N N N
M F
N N
N N N
During the observation period from the time of the lumbar puncture and MRI to the recent clinical evaluation of the patients (median 27 months, range 16-38 months), 13 patients developed CDMS. They all belonged to the subgroup of patients with abnormal MRI at onset. The relation between the development of CDMS and the results of MRI is statistically significant (p
