Relapsing demyelinating CNS disease in a Korean pediatric population: Multiple sclerosis versus neuromyelitis optica

Subjects and clinical data

Between 1994 and 2008, patients with relapsing CNS demyelinating episodes separated by intervals of at least 1 month were recruited at Seoul National University Children’s Hospital. Patients with evidence of cerebral vasculitis, granulomatous disease, or presumed infectious encephalitis were excluded. Definitions of the MS, NMO, recurrent or multiphasic ADEM and the clinically isolated syndromes (CIS) were used according to the consensus definitions. CIS were further differentiated as either polyfocal or monofocal, according to the number of CNS lesions involved that were relevant to clinical findings. If encephalopathy, defined as when a definite behavioral change and alteration in consciousness were present, was not associated, a diagnosis of polyfocal CIS was preferred to ADEM. In total, 21 patients were included and classified as having pediatric MS (n = 18) and pediatric NMO (n = 3). No case of recurrent or multiphasic ADEM was diagnosed during the same period. A retrospective analysis was performed, focusing on a description of clinical and MRI features at the initial episode and relapse, clinical course, the presence of AQP4 Ab, and the treatment response to interferon β-1b (Betaferon).

Cerebrospinal fluid assay

Polyacrylamide gel electrophoresis with isoelectric focusing was conducted and silver staining was used to detect oligoclonal bands.

AQP4 Ab Assay

AQP4 Ab assay was performed in three patients with NMO. Three additional patients, who did not meet the NMO criteria but who showed selective involvement of the optic nerve and spinal cord, were also included. The AQP4 assay was conducted at the Brain Research Institute, Niigata University, Niigata, Japan. Detailed methods for anti-AQP4 Ab detection are described in a previous report. ¹³

MRI data

Brain MRIs were performed using a 1.0- or 1.5-tesla system with axial T₂-weighted and fluid-attenuated inversion recovery sequences. Gadolinium enhancement was routinely conducted in all patients. Follow-up brain MRIs were performed either at the time of clinical relapses or 3 months after initial episode if no clinical relapse occurred during that period. Spine MRIs were performed and repeated at the same intervals as brain MRIs in a subset of patients who showed relevant clinical symptoms to the spinal cord. MRI lesions at the initial episode and relapse were defined and classified according to McDonald MRI subcriteria for dissemination in space. ¹⁴ Two MRIs (Case 6, Case 9) at the initial episode were excluded from characterizing McDonald MRI subcriteria, because only text reports of MRIs were available. All MRIs were reviewed by two radiologists (Cheon JE, Kim IO) who were blinded to the patient’s clinical history.

Treatment with Interferon β-1b

Interferon β-1b was administered to 14 patients (11 MS patients, three NMO patients) with the written consent of their parents. Patients who rejected the interferon treatment and experienced no further relapses after introduction of interferon β-1b into Korea (2002) were not treated. Treatment was started at 25% of full dose. The dose was escalated with careful monitoring of acute side effects. Most of the patients received 70–80% of the full adult dose as the titration endpoint. The efficacy of treatment was measured by the change in annualized relapse rate.

Pediatric MS (n = 18)

MRI features

The numbers of patients meeting McDonald MRI subcriteria for dissemination in space are provided in Table 1, and detailed individual findings are presented in Supplementary Table 2. Follow-up MRIs of two patients who showed no relapse, but met the MRI criteria of dissemination in space and time, were conducted 3 months after the initial episode and included in the analysis of MRI at relapse. Five patients (5/16, 31%) met the McDonald MRI criteria for dissemination in space at the initial episode, while nine patients (9/18, 50%) met the criteria at relapse.

OPEN IN VIEWER

Table 1.

Magnetic resonance imaging (MRI) features of 18 pediatric multiple sclerosis patients at the initial episode and relapse according to the McDonald MRI criteria

MRI features	Initial episode (n = 16)	Relapse (n = 18)
McDonald criteria	5 (31%)	9 (50%)
1 Gadolinium-enhancing lesion (A)	9 (56%)	11 (61%)
>9 T2-hyperintense lesions (B)	5 (31%)	13 (72%)
A or B	10 (63%)	14 (78%)
1 Infratentorial lesion	6 (38%)	10 (56%)
1 Juxtacortical lesion	7 (43%)	14 (78%)
3 Periventricular lesions	3 (18%)	6 (33%)

OPEN IN VIEWER

Table 2.

Interferon β-1b treatment data for 10 patients diagnosed as pediatric multiple sclerosis who continued the treatment at last follow-up

Case number	Onset (y)	Age at IFN start (y)	IFN duration (y)	Adverse reaction	AR before IFN	AR after IFN
4	5.6	7.2	3.4	H	1.2	0
5	5.7	7.6	3.5		1.1	0
7	6.5	7.3	3.2	H,F	3.5	0
8	6.6	7.9	2.3		2.2	1.3
10	6.6	7.9	2.3		2.2	1.3
12	8.0	12.6	4.5		0.7	0.4
13	8.4	8.6	1.4		5	0
14	8.6	9.9	0.3	H,L	3.2	0
17	10.1	10.5	5.6		10	0.2
18	13.6	13.9	1.4	H	7	0

IFN, interferon β-1b; H, headache; F, fever; L, liver enzyme elevations; AR, annualized relapse rate.

Pediatric NMO (n = 3)

One patient (Case 19, female) initially presented with bilateral optic neuritis at the age of 7 years and experienced subsequent multiple episodes of lower extremity weakness compatible with transverse myelitis. Brain MRI at the time of relapse revealed involvement of thalamus, corpus callosum, and periventricular area (Figure 1A). Spinal MRI first performed at relapse showed lengthy involvement of thoracic spinal segments (Figure 1B). Interferon was started at relapse, but discontinued because of frequent relapse and progressive clinical course. She died of pneumonia at the age of 14 years after disease duration of 7.2 years. The second patient (Case 20, male) initially presented with isolated transverse myelitis extending from his cervical to thoracic spinal cord at the age of 7.2 years. Brain MRI at the initial presentation revealed a subtle T₂-weighted high signal intensity in the area of the subcortical white matter of both temporo–occipital lobes. Three months later, he experienced bilateral optic neuritis. Interferon was started at relapse. During the 2.6-year follow-up period, no clinical relapse was observed with no disability and adverse events resulting from interferon treatment. The third patient (Case 21, female) experienced weakness in the left arm and leg associated with sensory change and visual disturbance in the left eye at the age of 11.7 years. Four clinical relapses occurred during 1.7 years: two episodes of optic neuritis and two of transverse myelitis. Brain and spinal MRI at the second relapse are presented in Figures 1Cand 1D. Interferon treatment was started after the fifth clinical episode. However, because of severe anemia (Hb 3.5 g/dl), we discontinued the treatment after 10 months of administration. Anemia did not begin to improve until 2 years after discontinuation despite prolonged use of a corticosteroid. Microscopic hematuria was detected on routine checkup 8 months later. Subsequent kidney biopsy confirmed lupus nephritis. Fluorescent antinuclear antibody, anti-double-stranded-DNA antibody, and complement C3 and C4 levels were all compatible with the diagnosis of systemic lupus erythematous. Her current visual system impairment corresponded to the Expanded Disability Status Scale score 3.

OPEN IN VIEWER

Figure 1.

Magnetic resonance images (MRIs) of a patient with neuromyelitis optica (Case 19; A, B). A, Axial fluid-attenuated inversion recovery image at relapse episode demonstrating high signal intensity periventricular and corpus callosal lesions. B, Sagittal T₂-weighted image at the second relapse demonstrating high signal intensity in the spinal cord over lengthy cervicothoracic spinal segments. MRIs of another patient with neuromyelitis optica (Case 21, C, D). C, Sagittal T₂-weighted image at the second relapse demonstrating high signal intensity in the long segment of the cervical and thoracic spinal cord. D, Axial fluid-attenuated inversion recovery image at the last relapse (fifth) episode demonstrating multifocal high signal intensity lesion in the subcortical and deep white matter of bilateral hemispheres.

AQP4 Ab Assay

The results of the AQP4 Ab assay and relevant clinical findings in six tested patients are summarized in Table 3. Two out of three NMO patients were positive for AQP4 Ab. Three patients in whom optic nerve and spinal cord were selectively involved, but who did not meet the definition of NMO, were all negative for AQP4 Ab.

OPEN IN VIEWER

Table 3.

Clinical profiles and anti-aquaporin-4 antibody results in six patients with selective involvement of the optic nerve and spinal cord

Case	Diagnosis	AQP4 Ab	Disease duration	Relapse numbers	MRI spinal cord involvement	IFN response	Current EDSS
19 (F)	NMO	+	7.2	13	++	Not effective	deceased
21 (F)	NMO	+	6.1	5	++	Not effective	3
20 (M)	NMO	–	2.6	2	++	Good	0
15 (F)	MS	–	1.6	4	++*	Good	4
7 (M)	MS	–	4.1	3	++*	Good	0
12 (M)	MS	–	8.9	5	+	Good	1

Pediatric MS

Three studies from different world regions examined the international consensus definitions following their proposal in 2007. ^15–17 All three studies focused on differentiation of MS from ADEM at the initial demyelinating event, suggesting more frequent conversion to MS from CIS than ADEM. Although we analyzed a selected group of patients who already suffered recurrent demyelinating events, none of them experienced ADEM at the initial and relapsing episode. ADEM was among the initial differential diagnosis in most of our patients presenting as polyfocal CIS at their initial and relapsing episodes. However, careful review of their history, focusing on the presence of encephalopathy, favors the diagnosis of polyfocal CIS rather than ADEM. Thus, our study could also support the evidence, although not prospectively followed, that current definitions that differentiate the initial demyelinating event of MS from ADEM depending on the presence of encephalopathy would be useful for prediction of relapses. Considering that two patients in our study (Case 1, Case 17; Supplementary Table 1) who were already confirmed with MS experienced subsequent ADEM attacks as their third demyelinating event, other clinical or biological markers explaining the unique pathophysiology of ADEM need to be further clarified.

In the present study, only one patient revealed a positive result for oligoclonal bands in cerebrospinal fluid. Even if the two further patients who showed an IgG index above 0.65 were added, the positive result (3/16, 19%) was still low considering most studies in the pediatric population reported positive results from 40–90% of patients. ¹⁸ As we did not use isoelectric focusing with immunoblotting for testing oligoclonal bands, ¹⁹ there may be a methodological limitation that contributes significantly to the low rate of positive detection. Moreover, similarly to Japanese adult MS patients where a low positive rate of oligoclonal band detection compared with MS patients of European descent was reported, ²⁰ ethnic differences might also exist in Asian pediatric MS patients. Thus, more sensitive testing methods (isoelectric focusing with immunoblotting) should be used to increase the reliability of interpretation of results in future studies in Asian pediatric populations.

Unsuccessful application of McDonald MRI criteria for dissemination in space in the pediatric age group has already been reported. ^21,22 The proportion of patients meeting McDonald MRI criteria for dissemination in space at the initial episode and relapse in our pediatric population was even lower than in other studies. Although information regarding gadolinium enhancement was available in all patients, this did not significantly influence the results because of the overall low positive rate in the other subcriteria, especially periventricular lesions. Specific conditions in our population including younger age onset, ethnic variation, and more frequent monofocal involvement of the optic nerve could be variables that caused the differences. Recently, modifications to the McDonald criteria in Asian adults was suggested, based on several Asian adult studies. ²³ Thus, differences in the clinical presentation need to also be considered when developing MRI criteria in the pediatric population.

Pediatric NMO and AQP4 assay

Among our 21 patients with recurrent demyelination, three patients met the current diagnostic criteria for NMO. These three patients, although the number is small, showed similar clinical and radiologic characteristics, female preponderance, association of AQP4 Ab and other autoimmune diseases, frequent relapse and poor prognosis, when compared with previous studies conducted in both adult and pediatric populations. ^8,24 All three patients showed asymptomatic, nonspecific brain lesions including corpus callosal, periventricular, and subcortical white matter lesions, although one patient (Case 19, Supplementary Table 2) met the McDonald MRI or Barkhof criteria.

When three patients among MS patients who showed selective involvement of their optic nerves and spinal cord but who did not met the criteria of the NMO were considered, six of 21 patients with recurrent demyelinating disease (6/21, 29%) revealed opticospinal involvement. To our knowledge, directly comparable studies in the pediatric population do not exist. This feature might be similar to studies in Japanese adults, in which opticospinal MS (OSMS) accounted for 15–40% of all MS cases. ²⁰ After identification of NMO-IgG or AQP4 Ab, most of Japanese OSMS patients are now considered to have an overlapping spectrum with NMO in Western populations. ²⁵ However, evidence regarding disagreement between clinical diagnosis of NMO, longitudinally extensive spinal cord lesions, and AQP4 autoimmunity was also presented in Japanese adult studies. ^26,27 Our two AQP4 Ab-positive patients were both female, and showed frequent relapse and early disabilities unresponsive to interferon treatment. It is noticeable that the other four patients, including one patient with NMO, who were negative for AQP4 Ab but showed heterogenous opticospinal involvement, seemed to experience milder clinical courses and responded favorably to interferon treatment despite the limitation of a variable follow-up period.

Although interferon treatment is generally regarded as ineffective for NMO, ²⁸ conflicting results in Japanese adult studies have also been reported. ^27,29 While Tanaka et al. reported no differences in annualized relapse rate between AQP4 Ab (+) and AQP4 Ab (-) NMO patients, ²⁹ Matsuoka et al. reported that eight out of 11 patients with AQP4 Ab negative OSMS with longitudinally extensive spinal cord lesions responded favorably to interferon β-1b. ²⁷ In addition, distinct immunological mechanisms have recently been suggested according to the status of AQP4 Ab, suggesting that a treatment response to interferon might be also affected by AQP4 Ab status. ³⁰ Although confirmatory evidence regarding interferon treatment could not be provided by our study due to small sample sizes, we have found heterogeneous opticospinal presentation in our pediatric population similar to that in Japanese adults. In addition, a possible link between AQP4 Ab status and disease course or interferon responsiveness was suspected. Clinical courses and treatment response to interferon need to be further characterized according to AQP4 Ab status. Prospective and longitudinally designed clinical studies equipped with both uniform brain and spine MRI protocols and AQP4 Ab results from the initial episode would help to confirm this clinical observation.