- Top of page
- What this paper adds
Aim Biomarkers such as autoantibodies, neopterin, and oligoclonal bands (OCBs) are increasingly used for the diagnosis of treatable inflammatory central nervous system (CNS) disorders. We investigated the correlation between the results of OCB testing and clinical diagnoses in a large contemporary cohort of children with a broad range of neurological conditions.
Method Cerebrospinal fluid (CSF) and serum from 200 children (94 females, 106 males; age range 2mo–15y 10mo, mean age 6y 9mo, SD ±4.9) who underwent CSF investigation for their neurological condition were tested for OCBs using isoelectric focusing.
Results The patients were divided into those with inflammatory (n=58) and non-inflammatory (n=142) CNS disorders. Intrathecal OCBs (OCBs restricted to the CSF) were found in 11 out of 58 (19%) of those with inflammatory CNS disorders compared with none of the 142 patients with non-inflammatory CNS disorders (p<0.001). Diseases associated with intrathecal OCB were multiple sclerosis, Rasmussen encephalitis, N-methyl-d-aspartate receptor (NMDAR) encephalitis, voltage-gated potassium channel (VGKC) encephalopathy, herpes (HSV) encephalitis, ‘other’ encephalitides, acute cerebellar ataxia, and aseptic meningitis. Mirrored OCBs (identical OCBs in the serum and CSF) were less specific but were still found in 14 out of 58 (24%) children with inflammatory CNS disorders compared with only 6 out of 142 (4%) children with non-inflammatory CNS disorders (p<0.001). Diseases associated with mirrored OCBs included acute disseminated encephalomyelitis (ADEM), VGKC encephalopathy, West syndrome, NMDAR encephalitis, ‘other’ encephalitides, polio-like illness, Rasmussen encephalitis, cerebral vasculitis, metachromatic leukodystrophy, and bacterial meningitis. Intrathecal OCBs and mirrored OCBs had a positive predictive value for inflammatory CNS disease of 1 (95% confidence interval [CI] 0.68–1) and 0.7 (95% CI 0.46–0.87) respectively.
Conclusion Intrathecal OCBs were restricted to patients with inflammatory CNS disorders. They are a useful, but non-specific, biomarker of CNS inflammation of multiple causes. Mirrored OCBs are less specific, but still support a possible inflammatory CNS disorder. The presence of either intrathecal or mirrored OCBs should raise suspicion of an inflammatory CNS disorder.
There is increasing interest in biomarkers that diagnose potentially treatable inflammatory central nervous system (CNS) disorders.1 Some serum and cerebrospinal fluid (CSF) autoantibodies are specific biomarkers associated with autoimmune syndromes such as N-methyl-d-aspartate receptor (NMDAR) encephalitis and voltage-gated potassium channel (VGKC) encephalopathy. Other biomarkers, including CSF neopterin and oligoclonal bands (OCBs), are less specific, but they are still useful markers of immune activation or inflammation in the CNS.2 Before the discovery of the NMDAR antibody biomarker, OCBs were used to support the hypothesis that encephalitis lethargica and immune-mediated chorea encephalopathy syndrome were autoimmune diseases; these syndromes have subsequently been shown to be NMDAR encephalitis.3–5 OCBs have also been useful in the identification of unusual variants of known autoimmune encephalopathy.6
OCBs are clones of immunoglobulin (typically IgG) that can be detected in the CSF and/or serum. The qualitative method of isoelectric focusing on agarose gels followed by immunoblotting is the accepted standard method for OCB detection. The IgG index (the ratio of CSF/serum IgG to CSF/serum albumin), a quantitative method of analysis, is a less sensitive test, and when elevated is suggestive of a CNS B-cell response. There are limited reports of use of the IgG index in paediatric studies, but in adult studies the IgG index has been found to be only rarely elevated in patients with multiple sclerosis who have negative OCBs.7–11 Alper et al.,7 in a paediatric cohort, found that the IgG index was less frequently elevated in those with acute disseminated encephalomyelitis (ADEM) than in those with multiple sclerosis.7 Interpretation of OCBs is dependent on the comparison of CSF and serum samples, with a number of patterns described (Fig. 1).8
Figure 1. Isoelectric focusing on agarose gels followed by immunoblotting. Lanes 1 and 2: no oligoclonal bands (OCBs) detected in cerebrospinal fluid (CSF) or serum. Lanes 3 and 4: OCB detected in CSF but not in serum indicative of intrathecal immunoglobulin G (IgG) production. Lanes 5 and 6: identical OCB detected in both CSF and serum suggestive of an abnormal CSF-blood barrier (mirrored pattern). Lanes 7 and 8: combination of mirrored and intrathecal IgG production, a pattern not observed in this study but observed in subacute sclerosing panencephalitis.
Download figure to PowerPoint
When OCBs are present in the CSF and absent from the serum, this is termed an ‘intrathecal pattern’ and is indicative of local antibody production and hence a humoral immune response within the CNS. When identical clones of IgG are present in the CSF and serum, this is termed a ‘mirrored pattern’ and suggests that IgG has entered the CNS from the systemic circulation, such as occurs in blood–brain barrier damage.
A combination of mirrored and intrathecal patterns can occur, in which identical clones of IgG are present in the CSF and serum; however, the CSF also contains additional OCBs. This is consistent with both CNS contamination and local (intrathecal) IgG synthesis. Intrathecal OCBs are accepted to be of greater clinical significance than mirrored OCBs in the diagnosis of inflammatory CNS disorders.10
Data pertaining to OCBs in paediatric patients are largely restricted to specific diseases, for example multiple sclerosis,12 ADEM,13 and opsoclonus–myoclonus syndrome.14 In 1986, Kostulas et al.9 reported OCB findings in a paediatric cohort with varying neurological diagnoses. They found OCBs most commonly in patients with inflammatory disorders; however, they also reported finding OCBs in children with presumed non-inflammatory disorders such as epilepsy and migraine. In recent years, the ability to diagnose disorders of inflammatory aetiology has greatly improved. We studied the diagnostic associations and sensitivity/specificity of intrathecal and mirrored OCB in a contemporary cohort of children with neurological diseases.
- Top of page
- What this paper adds
We identified 205 paediatric neurology patients who had OCB testing between January 2006 and February 2010. During that time, OCB testing was routinely requested for all paediatric neurology patients at the Children’s Hospital at Westmead undergoing CSF analysis. This study was part of ethically approved studies of inflammatory and autoimmune disorders of the CNS at the hospital. We excluded five patients in whom paired CSF and serum samples or clinical information was unavailable. Patients were aged 2 months to 15 years 10 months, with a mean age of 6 years 9 months SD ±4.9. There were 94 females and 106 males. All CSF and sera specimens were analysed by isoelectric focusing at a single laboratory. In the case of those patients with acquired disorders, analysis was performed on samples taken during the acute period of illness, and in the first episode if the disorder was relapsing. The study also included patients with chronic illnesses who were undergoing CSF analysis in an ‘elective’ manner. Patient case notes were reviewed and patients were categorized by pathophysiology (inflammatory vs non-inflammatory CNS disease), disease group, and diagnosis (Table I). Case note review and patient categorization were performed by a neurology registrar (AJS), with verification provided by a paediatric neurologist (RCD). Assigned diagnoses were based on those documented in the case notes by the treating paediatric neurologist. The inflammatory CNS group (n=58; 29%) included patients with diagnoses with an established infectious, autoimmune, immune-mediated, or inflammatory cause. The term ‘inflammatory CNS’ will be used to describe infectious, autoimmune, inflammatory, or immune-mediated CNS disorders. The non-inflammatory CNS group (n=142; 71%) included patients with diagnoses for which there was no established inflammatory pathophysiology and patients in whom we were unable to assign a diagnosis.
Table I. Disease groups and diagnoses
|Disease group (n)||Diagnoses (n)|
|Non-inflammatory CNS (n=142)|
| Epilepsy (n=58)||Electroclinical syndromes: West syndrome (n=6), Dravet syndrome (n=1), febrile seizures plus (n=1), epilepsy with myoclonic astatic/atonic seizures (n=2), juvenile absence epilepsy (n=1)|
|Structural or metabolic: perinatal brain injury (n=1), traumatic brain injury (n=1), hypoglycaemia (n=1), malformation of cortical development (n=4), previous CNS infection (n=3)|
|Genetic: PCDH19 mutation (n=1), SCN1A (n=2)|
|Other or unknown (n=34)|
| Static encephalopathy (n=15)||Cerebral palsy (n=3), developmental delay (n=4), chromosomal disorders (n=6), unknown (n=2)|
| Progressive genetic or metabolic (n=8)||Metachromic leukodystrophy (n=1), Menkes disease (n=1), mitochondrial/probable mitochondrial (n=2), hereditary spastic paraplegia (n=2), unknown (n=2)|
| Other (n=61)||Functional disorders (n=10), movement disorders (n=10), Guillain–Barré syndrome (n=5), headache disorders (n=4), stroke (n=4), miscellaneousa (n=13), unknown (n=15)|
|Inflammatory CNS (n=58)|
| Demyelinating (n=26)||MS (n=6), ADEM (n=11), clinically isolated syndrome (n=9): optic neuritis (n=3), transverse myelitis (n=3), other (n=3)|
| Infection-mediated (n=11)||Bacterial meningitis (n=1), aseptic meningitis (n=1), ‘other’ encephalitides (n=6), acute necrotizing encephalopathy (n=2), HSV encephalitis (n=1)|
| Autoantibody-associated (n=7)||NMDAR encephalitis (n=3), VGKC encephalopathy (n=3), basal ganglia encephalitis (n=1)|
| Other immune-mediated (n=14)||Rasmussen encephalitis (n=3), acute cerebellar ataxia (n=4), Polio-like illness (n=2), cerebral vasculitis (n=1), opsoclonus–myoclonus syndrome (n=2), Sydenham chorea (n=1), Aicardi–Goutières syndrome (n=1)|
We subdivided the inflammatory CNS group into demyelinating diseases (n=26), infection-mediated diseases (n=11), autoantibody-associated diseases (n=7), and ‘other’ immune-mediated diseases (n=14). Diagnoses within these subgroups are presented in Table I. Within the infection-mediated group, six patients were categorized as having ‘other’ encephalitides. These patients had evidence of encephalitis (acute or chronic) fulfilling criteria for encephalitis but without definite evidence of a specific infectious agent.15 The non-inflammatory CNS group included patients with very diverse diagnoses (Table I) along with patients in whom the diagnosis was unknown (n=15). The two-tailed Fisher’s exact test was used to calculate p-values, and a p-value <0.05 was considered significant; 95% confidence intervals (CI) are presented for positive predictive value and negative predictive value.
- Top of page
- What this paper adds
Intrathecal OCBs were detected in patients with inflammatory CNS disorders (11/58; 19%), but not in those with non-inflammatory CNS disorders (0/142; 0%; p<0.001; Table II). Diagnoses associated with intrathecal OCBs were multiple sclerosis (n=3), Rasmussen encephalitides (n=2), NMDAR encephalitis (n=1), VGKC encephalopathy (n=1), acute cerebellar ataxia (n=1), aseptic meningitis (n=1), herpes simplex virus (HSV) encephalitis (n=1), and ‘other’ encephalitis (n=1; Table II). Mirrored OCBs were found more frequently in those with inflammatory CNS disorders (14/58, 24%) than in those with non-inflammatory CNS disorders (6/142; 4%; p<0.001; Table II). Mirrored OCBs were detected in patients with ADEM (n=4), VGKC encephalopathy (n=1), NMDAR encephalitis (n=2), basal ganglia encephalitis (n=1), ‘other’ encephalitides (n=1), polio-like illness (n=2), Rasmussen encephalitis (n=1), varicella zoster-associated cerebral vasculitis (n=1), bacterial meningitis (n=1), West syndrome (n=2), and metachromatic leukodystrophy (n=1), and in several patients with unknown diagnoses (n=3). No patient showed a combination of mirrored and intrathecal patterns.
Table II. Clinical correlation of intrathecal and mirrored oligoclonal bands (OCBs)
| ||OCB result|
|Disease groups and diagnoses||Intrathecal, n (%)||Mirrored, n (%)|
|Total||11/200 (6)||20/200 (10)|
|Non-inflammatory CNS||0/142 (0)||6/142 (4)|
| Epilepsy||0/58 (0)||2/58 (3)|
| Static encephalopathy||0/15 (0)||0/15 (0)|
| Progressive genetic or metabolic||0/8 (0)||1/8 (13)|
| Other||0/61 (0)||3/61 (5)|
|Inflammatory CNS||11/58 (19)||14/58 (24)|
| Demyelinating||3/26 (12)||4/26 (15)|
| ADEM||0/11 (0)||4/11 (36)|
| Clinically isolated syndrome||0/9 (0)||0/9 (0)|
| Multiple sclerosis||3/6 (50)||0/6 (0)|
| Infection-mediated||3/11 (27)||2/11 (18)|
| Bacterial meningitis||0/1||1/1|
| Aseptic meningitis||1/1||0/1|
| HSV encephalitis||1/1||0/1|
| ‘Other’ encephalitides||1/6||1/6|
| Acute necrotizing encephalopathy||0/2||0/2|
| Autoantibody-associated||2/7 (29)||4/7 (57)|
| NMDAR encephalitis||1/3||2/3|
| VGKC encephalopathy||1/3||1/3|
| Basal ganglia encephalitis||0/1||1/1|
| Other immune-mediated||3/14 (21)||4/14 (29)|
| Rasmussen encephalitis||2/3||1/3|
| Acute cerebellar ataxia||1/4||0/4|
| Polio-like illness||0/2||2/2|
| Opsoclonus–myoclonus syndrome||0/2||0/2|
| Varicella zoster vasculitis||0/1||1/1|
| Aicardi–Goutières syndrome||0/1||0/1|
| Sydenham chorea||0/1||0/1|
Intrathecal OCBs had a sensitivity of 0.19 (95% CI 0.10–0.32), a specificity of 1 (95% CI 0.97–1), a positive predictive value of 1 (95% CI 0.68–1), and a negative predictive value of 0.75 (95% CI 0.68–0.81) for the presence of inflammatory CNS disease (Table II). Mirrored OCBs had a sensitivity of 0.24 (95% CI 0.14–0.37), a specificity of 0.96 (95% CI 0.91–0.98), a positive predictive value of 0.7 (95% CI 0.46–0.87), and a negative predictive value of 0.76 (95% CI 0.68–0.82; Table II).
Those with autoantibody-associated disorders were likely to have either intrathecal or mirrored OCBs (6/7; 86%). In both patients with NMDAR encephalitis whose initial samples showed mirrored OCBs, CSF samples taken 59 and 70 days after the first samples were found to contain intrathecal OCBs. Among those with demyelinating disorders, intrathecal OCBs were detected only in patients with multiple sclerosis (3/6; 50%), whereas mirrored OCBs were detected only in those patients with ADEM (4/11; 36%). All three patients with Rasmussen encephalitis had either intrathecal or mirrored OCBs. Patients with infection-mediated disorders (n=11) had either intrathecal (3/11; 27%) or mirrored OCBs (2/11; 18%).
- Top of page
- What this paper adds
We sought to investigate the utility of OCBs in a contemporary group of child neurology patients and make diagnostic associations. We found intrathecal OCBs only in patients with evidence of inflammatory CNS disease; however, a variety of autoantibody-associated, demyelinating, infection-mediated, and other immune-mediated diseases were associated with intrathecal OCBs. Mirrored OCBs were less specific than intrathecal OCBs for CNS inflammation, but were still more likely to be found in patients with inflammatory CNS disease than in those with non-inflammatory CNS disease (24% vs 4%; p<0.001) and therefore can also be of clinical utility. Mirrored OCBs imply the presence of clonal IgG in both CSF and serum. Indeed, many of the autoantibody biomarkers that are useful in CNS inflammation, such as antibodies against neuromyelitis optica/aquaporin-4, VGKC, leucine-rich glioma-inactivated 1, and myelin oligodendrocyte glycoprotein are measured in the serum, rather than in the CSF. In the case of many autoantibody-associated CNS disorders, the production of autoantibody might be first triggered in the periphery, rather than in the CNS, and the importance of serum versus CSF antibody production is a central emerging theme in neuroimmunology.16,17 For these reasons, we believe that mirrored OCBs should not be ignored but may be an important biomarker in inflammatory CNS disorders. Many of the patients in this report with mirrored OCBs had autoantibody-associated disorders, such as NMDAR encephalitis and VGKC encephalopathy. Interestingly, we noted that the patients with NMDAR encephalitis who initially had mirrored OCBs subsequently developed intrathecal OCBs, suggesting that patients initially have a systemic autoimmune response, which subsequently becomes localized intrathecally. No patient was found to have a combination of both a mirrored and intrathecal OCB pattern, suggesting that this may be an uncommon finding in paediatric populations, or a pattern that develops in chronic disease, such as in subacute sclerosing panencephalitis.
Specific diseases associated with OCBs were generally consistent with findings previously reported by others. Among our patients with demyelinating disorders (n=26), intrathecal OCBs were detected only in patients with multiple sclerosis, and mirrored OCBs were found only in patients with ADEM. Our numbers were small; however, studies specifically of paediatric patients with demyelinating diseases have also reported a higher frequency of intrathecal OCBs in children with multiple sclerosis than with ADEM or CIS.7,12,13,18 The presence of mirrored OCBs in ADEM has also been reported in other studies.19,20 Intrathecal OCBs were also found in patients with Rasmussen encephalitis and acute cerebellar ataxia, as previously reported.21,22
Although OCBs can be useful, we failed to detect OCBs in many patients with recognized inflammatory CNS diseases. The absence of OCBs should not preclude consideration of an inflammatory CNS process, or testing for specific autoantibodies.
There is emerging evidence that autoantibodies against VGKC and NMDAR are specific in autoimmune encephalopathy and it could be argued that these specific biomarkers will increasingly replace non-specific markers of inflammation or immune activation, such as OCBs and CSF neopterin.1,2 However, non-specific biomarkers are likely continue to be of clinical utility as they can alert the clinician to atypical variants of autoimmune encephalopathy or to potential autoimmune diseases yet to be fully defined.6 Additionally, specific autoantibody tests are less widely available than OCBs and may have a longer processing time; given that early diagnosis and treatment can improve outcomes, OCBs are likely to continue to play a useful role.
Assessing a diagnostic test such as OCBs ideally requires comparison with a criterion standard, but unfortunately there is no practical criterion standard test for CNS inflammation as brain biopsy is too invasive. The group with CNS inflammatory disorders in this study is a heterogeneous group in which the precision of diagnosis and the certainty regarding the inflammatory basis of the disease varies substantially. A prospective study with an extended follow-up time would allow for collection of all pertinent data (such as auxiliary evidence of inflammation) and would increase the level of accuracy for the ‘less precise’ diagnoses and would perhaps reduce the numbers of patients in whom the diagnosis is unknown.