Article Text

Download PDFPDF

Original research
Oligoclonal IgG bands in chronic inflammatory polyradiculoneuropathies
  1. Marta Ruiz1,
  2. Marco Puthenparampil1,
  3. Marta Campagnolo1,
  4. Francesca Castellani1,
  5. Alessandro Salvalaggio1,2,
  6. Susanna Ruggero1,
  7. Elisabetta Toffanin1,
  8. Mario Cacciavillani3,
  9. Paolo Gallo1,
  10. Diego Franciotta4,
  11. Chiara Briani1
  1. 1 Department of Neurosciences (DNS), University of Padova, Padova, Italy
  2. 2 Padova Neuroscience Center (PNC), Padova, Italy
  3. 3 EMG Lab, CEMES, Synlab, Padova, Italy
  4. 4 IRCCS, Ospedale Policlinico San Martino, Genova, Italy
  1. Correspondence to Dr Chiara Briani, University of Padua, 35128 Padova, Veneto, Italy; chiara.briani{at}


Background Cerebrospinal fluid (CSF) albumincytologic dissociation represents a supportive diagnostic criterion of chronic inflammatory demyelinating polyradiculoneuropathy (CIDP).Few studies have investigated possible systemic or intrathecal humoral immune response activation in CIDP.

Aim of our study was to investigate whether the search of oligoclonal IgG bands (OCBs) might provide additional data helpful in CIDP diagnostic work-up.

Methods Forty-eight consecutive patients with CIDP (34 men, mean age 59.4, range 16–83) were recruited. CSF analysis included nephelometric measurement of albumin and IgG concentrations, calculation of QALB, QAlbLIM and intrathecal IgG synthesis, and OCBs detection with isoelectric focusing. Data were compared with those from CSF and serum of 32 patients with Guillain-Barré syndrome (GBS), 18 patients with anti-myelin associated glycoprotein (MAG) antibody neuropathy, 4 patients with multifocal motor neuropathy and 32 patients with non-inflammatory neuropathies (NINPs).

Results Patients with CIDP and anti-MAG antibody neuropathy had significantly higher CSF albumin concentrations and QALB values than NINPs (p=0.0003 and p=0.0095, respectively). A total of 9 (19%) patients with CIDP presented identical serum and CSF OCBs (‘mirror pattern’) versus 3 patients (16.6%) with anti-MAG antibody neuropathy, 13 patients (40.6%) with GBS and 12.5% patients with NINPs. Only one patient with CIDP showed unique-to-CSF OCBs. First-line therapy was effective in 80.4% of patients with CIDP, irrespective of CSF findings.

Conclusions Compared with NINP, CIDP, GBS and anti-MAG antibody neuropathies had a significantly increased CSF protein and blood–spinal nerve root barrier damage. Intrathecal humoral immune response is rare in our patients with CIDP. Systemic oligoclonal activation is more frequent, but not significantly different from what was detected in the control groups.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information. Data (deidentified participant data) are available upon reasonable request to the corresponding author.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Chronic inflammatory demyelinating polyradiculoneuropathy (CIDP) is an acquired immune-mediated inflammatory disorder that may have a chronic progressive or relapsing course.1 Diagnosis rests on a combination of clinical and neurophysiological features, which are typical of acquired primary peripheral demyelination.2 Elevated cerebrospinal fluid (CSF) protein content with leucocyte count <10/mm3, defined as albumincytologic dissociation, represents a marker of blood–spinal nerve root barrier (B–SNR-B) permeability that is however not specific of CIDP,3 and is thus considered only as a supportive diagnostic criterion.2 Recently, Franciotta et al suggested to replace the albumincytologic dissociation with the albumin quotient (QALB=CSF albumin/serum albumin), considered a more accurate index of B–SNR-B permeability.4 Moreover, and rather unexpectedly, CSF-restricted oligoclonal IgG bands (OCBs), a common finding in immunoinflammatory central nervous system disorders, especially multiple sclerosis, have also been detected in some patients with chronic inflammatory neuropathies,5–8 thus suggesting an activation of an intrathecal immune response concurrent with the well-documented systemic reaction. In addition to the paucity of studies regarding this issue, suboptimal (eg, agarose electrophoresis) and heterogeneous techniques have been employed, with the population studied mainly represented by patients with Guillain-Barré syndrome (GBS) and only a small cohort of patients with CIDP.8 A few studies have also analysed CSF single IgG bands,9–13 which are more common in central nervous system inflammatory disorders, and may become OCBs at follow-up, especially in multiple sclerosis.10 12 CSF single IgG bands have been reported also in a small percentage of patients with peripheral nervous system diseases.9 10 12 However, their significance is unclear, and it is unknown whether they might reflect a response to a persisting antigenic stimulation. Finally, the mirror pattern, namely the presence of identical OCBs in both CSF and serum, which provides evidence of systemic immune activation in the absence of intrathecal IgG synthesis, has been described also in patients with inflammatory neuropathies.5–8

The aim of our study was to investigate the CSF profile, which includes the currently determined laboratory parameters and the less frequently investigated OCBs, in a cohort of consecutive patients with CIDP observed at the time of diagnosis, in order to identify possible features that might be helpful for diagnostic and/or prognostic purposes.

Data were compared with those from patients with other immune-mediated neuropathies, namely GBS and anti-MAG antibody neuropathy, and from patients with other non-inflammatory neuropathies (NINPs), enrolled in the same period of time.


Forty-eight consecutive patients (34 men, mean age 59.4±15.5 years, range 16–83 years), evaluated from September 2004 to September 2019 at the neuropathy centre in Padova, diagnosed with definite CIDP according to the European Federation of Neurological Societies/Peripheral Nerve Society guidelines,2 were considered. Among these patients, 33 had classical CIDP and 15 atypical variants (4 multifocal acquired demyelinating sensory and motor (MADSAM) neuropathy, 3 pure motor CIDP, 3 pure sensory CIDP, 3 distal acquired demyelinating symmetric neuropathies and 2 focal CIDP). Clinical assessment and disability were scored according to the Inflammatory Neuropathy Cause and Treatment (INCAT) Scale14 and the Inflammatory Rasch-Built Overall Disability Scale (I-RODS)15 at diagnosis and at follow-up evaluations. The patterns of clinical course were classified as previously described.16 With regard to the disease course, the majority (86%) of patients had a progressive course, 5 (10%) patients had a relapsing-remitting CIDP and 2 (4%) patients had a monophasic CIDP. Response to treatment was defined as improvement of at least one point in one disability scale (INCAT or I-RODS) after the first 6 months of therapy. Clinical worsening of one or more points in the clinical scales after 6 months of treatment was necessary to define therapy inefficacy. Participants who did not improve by one point after 6 months received alternative treatment.

CSF analysis was performed, after written informed consent, at first admission at our hospital before starting any therapy. As controls we considered 86 consecutive patients, 32 patients with GBS, 18 with anti-MAG antibodies neuropathy, 4 with multifocal motor neuropathy and 32 patients with other NINPs (19 with axonal idiopathic neuropathy, 5 with sensory neuronopathy with negative anti-Hu antibodies, 1 with hereditary neuropathy, 1 with diabetic neuropathy, 1 with neurolymphomatosis, 1 with small fibre neuropathy, 4 with other demyelinating neuropathies; mean age at recruitment 61.9±13.2 years, range 31–83 years, mean disease duration 2.3±2.3 years, range 0.5–12 years), who underwent CSF analysis at our hospital during the same time period as the 48 patients with CIDP. Baseline characteristics of the patients are summarised in table 1.

Table 1

Demographic characteristics of the study population

CSF analysis

CSF was collected by non-traumatic lumbar puncture in the sitting position. Routine examination on paired CSF and serum specimens included: CSF red cell count, CSF white blood cell and differential cell count, CSF concentration of total proteins, albumin and IgG determination and search for OCBs, which was done by means of agarose isoelectric focusing (IEF) followed by transfer to nitrocellulose membrane, IgG specific immunofixation, amplification with avidin–biotin and peroxidase staining.17 A CSF protein concentration greater than 45 mg/dL was defined as abnormally increased. We also assessed the CSF/serum albumin ratio (QALB). For the functional assessment of B–SNR-B permeability, along with the QAlbLIM (calculated as (age/15)+4) as cut-off. B–SNR-B damage was classified as absent when QALB<QAlbLIM was maximum, as mild B–SNR-B damage if QALB/QAlbLIM was lower than 2.0, as moderate B–SNR-B damage if QALB /QAlbLIM was between 2.0 and 5.0 and as severe when over 5.0.18 19

The interpretation of OCBs pattern was performed in line with the current guidelines for CSF study in neuroinflammatory diseases of the central nervous system.20 For the purpose of this study, we considered the following pathological patterns: (a) CSF-restricted OCBs (pattern 2: presence of intrathecal IgG synthesis); (b) identical OCBs in both serum and CSF (pattern 4: mirror pattern, systemic OCB production, without intrathecal IgG synthesis) and (c) the identification of identical OCBs in serum and CSF, with regular and periodic spacing and decreasing intensity (pattern 5: paraproteinemic pattern, presence of monoclonal gammopathy).20

Statistical methods

Descriptive statistic was reported for the entire sample of patients with CIDP and for control group. Categorical variables were described by using percentages, while continuous variables were described by using means, medians and ranges. Demographic, clinical and CSF features, treatment response and disability level were compared between different subgroups of patients with the χ2 test for categorical variables and with the t-test or the Kruskal-Wallis test for continuous variables. Post hoc analysis was assessed by ANOVA with Bonferroni correction for multiple comparisons. Correlation analysis was calculated with Pearson rank correlation. The significance level was set at 0.05.


At spinal tap, patients with CIDP had mean disease duration of 2.9±4.6 years, a median INCAT disability score of 2.0 (range 0–7) and a median I-RODS value of 14 (mean 16.1, range 2–38). Regarding comorbidities, among patients with CIDP, 4/48 (8.3%) patients had multiple myeloma (1 smouldering IgAk, 1 light chain myeloma, 2 IgGk) and 6/48 (12.5%) patients had a monoclonal gammopathy of undetermined significance (MGUS) (2 IgMλ, 1 IgMk, 2 IgGk, 1 double MGUS IgGk+IgAλ). Patients with CIDP and concurrent MGUS were negative for anti-MAG antibodies. Among patients with NINPs, 2/32 (6.2%) patients, both with chronic lymphocytic leukaemia, had an IgG MGUS (1 IgGk, 1 IgGλ). Patients with CIDP, anti-MAG antibody neuropathy and NINPs did not differ in demographic characteristics (table 1).

The main CSF data are reported in table 2.

Table 2

CSF data in patients with CIDP and control groups

Compared with NINPs, CSF protein concentration was increased in both patients with CIDP (35.1±10.0 mg/dL vs 63.4±58.2 mg/dL, p=0.032) and GBS (76.8±44.5 mg/dL, p=0.0013). In addition, increased CSF protein was more frequently observed in CIDP (26/48 patients, 54.2%, p=0.0009), GBS (23/32 patients, 71.9%, p<0.0001) and anti-MAG (10/18 patients, 55.5%, p=0.0101) groups than in NINPs (5/32 patients, 15.6%).

Both patients with CIDP and GBS had significantly higher CSF albumin concentrations (p=0.044 and p<0.0001, respectively) and QALB (p=0.0044 and p<0.0001, respectively) than NINPs (table 2). B–SNR-B damage was observed in 26/48 (54.2%) patients with CIDP (18 mild, 6 moderate, 2 severe), in 24/32 (75.0%) patients with GBS (10 mild, 13 moderate, 1 severe), in 5/32 patients (15.6%) with NINPs (mild) and in 12/18 patients (66.7%) with anti-MAG antibody neuropathy (8 mild, 4 moderate). No difference between typical and atypical CIDP was observed for CSF protein concentration (p=0.27), QALB (p=0.13) and QALB ratio (p=0.29), as well as for the frequency of BNB damage (p=0.77). A strong correlation between CSF protein concentration and QALB ratio was observed in all groups: CIDP (r: 0.89, r2: 0.79, p<0.0001), GBS (r: 0.71, r2: 0.50, p<0.0001), anti-MAG antibody neuropathy (r: 0.92, r2: 0.84, p<0.0001) and NINPs (r: 0.82, r2: 0.68, p<0.0001) (figure 1).

Figure 1

Correlation between CSF protein concentration and QALB in the four groups of patients. CIDP, chronic inflammatory demyelinating polyradiculoneuropathy; CSF, cerebrospinal fluid; GBS, Guillain-Barré syndrome; NINPs, non-inflammatory neuropathies.

A pattern 4 (mirror pattern) was observed in 9 patients with CIDP (9/48, 19%) (table 3) (figure 2B). Among them, we identified two patients, affected with light chain myeloma and smouldering IgAk myeloma, respectively, in whom a paraproteinemic pattern was not present, since the free light chains (produced in light chain myeloma) are detected only in urine. Another patient with a double IgGk+IgAλ MGUS presented with a mirror pattern, whereas the typical ladder-like paraproteinemic pattern (linked to the IgG monoclonal component) was not detectable, possibly due to the small amount of the paraprotein that may be overlooked on IEF.

Table 3

Mirror pattern in the four subgroups of patients

Figure 2

A: pattern V+II; B: pattern IV. Isoelectric focusing from paired serum (S) and cerebrospinal fluid (CSF) specimens of two patients with CIDP. Double arrowhead indicates present in both S and CSF, while single arrowhead indicates CSF-restricted IgG OCBs. On the left, ‘a typical monoclonal band appearance in both CSF and S characterizes pattern 5. However, additional CSF-restricted oligoclonal bands are clearly shown, defining a 5+2 pattern. On the right, identical S and IgGOB (mirror pattern 4) are shown. CIDP, chronic inflammatory demyelinating polyradiculoneuropathy. IgGOB, IgG oligoclonal bands.

Three patients (3/48, 6%) had a pattern 5 (2 IgGk myeloma, 1 IgGk MGUS). In one patient (1/48, 2%) with CIDP we observed unique-to-CSF OCBs associated with a paraproteinemic pattern (figure 2A). This patient had concomitant MGUS IgGk and diabetes, while no central nervous system immune-mediated comorbidities were observed. Brain MRI was unremarkable, and paranodal antibodies to NF155, CNTN1 and CASPR1 were negative. Patients with IgM MGUS (2 IgMλ, 1 IgMk) disclosed no abnormalities. A pattern 4 was identified in 3/18 (16.6%) patients with anti-MAG antibody neuropathy, and in 13/32 (40.6%) patients with GBS. The presence of pattern 4 increased the risk of CIDP or GBS diagnosis (O.R. 16.2, 95% CI 6.2 to 39.2, p<0.0001). Finally, IEF from patients with NINPs showed pattern 4 in 4/32 patients (12.5%, 2 axonal idiopathic neuropathy, 1 sensory neuronopathy, 1 hereditary neuropathy) and pattern 5 in previously described 2/32 (6.2%) patients, 1 axonal idiopathic neuropathy with a MGUS IgGk and 1 neurolymphomatosis with a MGUS IgGλ, both associated chronic lymphocytic leukaemia.21 Only one patient (3.1%) with axonal neuropathy had OCBs in the CSF (pattern 2). The patient had undergone spinal MRI, which revealed mild disc arthrosis.

Response to treatment

Patients with CIDP were followed up for 2.4±3.4 years (median 2.4, range 0–14.5). A total of 41 (85.4%) patients underwent first-line immune-modulatory therapy during the time of the study. Among them, 35 were treated with intravenous immunoglobulins (IVIg), 4 with steroids and 1 with plasma exchange (PE). A total of 33 (80.4%) patients showed clinical improvement of at least one point in one disability scale in the first 6 months of therapy and were considered responders (28 to IVIg, 1 to PE, 4 to steroids). The remaining 8 (19.5%) patients failed to respond to first-line treatment. Among them, three underwent a second-line treatment with benefit.

There was no statistically significant difference in the response to treatment among the patients with CIDP with normal IEF and those with other (mirror, paraproteinemic or unique-to-CSF OCBs) patterns.

Patients with CIDP with a moderate B–SNR-B damage had a significantly younger disease onset than patients with CIDP with mild (p=0.019) or no B–SNR-B damage (p=0.011).


In the diagnostic work-up of patients with CIDP, elevated CSF protein levels represents only a supportive diagnostic criterion in the presence of minimal mandatory clinical and electrodiagnostic data, due to its low specificity.2 However, increased CSF proteins may be observed during daily clinical practice and may often be misleading, also because consensus on standard CSF total protein upper reference limit is still lacking.22 23

The results of our study in 134 consecutive patients with different type of neuropathy showed that CSF protein and B–SNR-B damage are common findings in immune-mediated neuropathies, different from NINPs. It has been suggested that differences in the pattern of B–SNR-B dysfunction may contribute to the diverse distribution of demyelination and to clinical phenotypes, namely typical versus atypical CIDP.24–26 B–SNR-B breakdown has also been reported by biopsy specimens.27 28 Fine ultrastructural changes, including fenestration of endoneurial microvessels and disappearance of tight junctions, have been observed in sural nerve biopsies from patients with CIDP.25 These findings suggest that B–SNR-B may be the entry point of pathogenic T cells and other soluble factors with increased permeability playing a possible role. Upregulation of serum proinflammatory cytokines or direct attack of antibodies to B–SNR-B components have been proposed as possible pathogenic mechanisms.27 28 A recent study29 reported that patients with typical CIDP, in whom distal and proximal segments are predominantly involved, show unremarkable pathological features in sural nerve biopsies. On the contrary, patients with atypical forms, such as MADSAM or pure sensory CIDP, are characterised by focal lesions in the nerve trunks sparing the sites where B–SNR-B is deficient in typical CIDP.

As for anti-MAG antibody neuropathy, CSF protein can be increased in patients with this disease, with normal cell count.30 The pathogenic mechanisms involved in this abnormality are unclear, in both anti-MAG IgM-associated and in other non-anti-MAG IgM-associated neuropathies, where albuminocytological dissociation can occur in CIDP-like phenotypes.31 The possibility of a damaged root–CSF barrier in anti-MAG antibody neuropathy has already been proposed.32 ,33

Regarding the IEF findings, previous studies with suboptimal and heterogeneous techniques reported CSF oligoclonal bands, not detected with IgG-specific procedures, in numerous patients with CIDP5–8 (19/47 patients with acute idiopathic polyneuropathy and 14/15 of those with chronic relapsing polyneuropathy5; 13/16 patients with GBS of whom 10 in serum only and 3 in serum and CSF7; 68% of 72 patients with GBS; 79% of 24 patients with CIDP8; in 197 control patients with headache or psychiatric disorder, a weak oligoclonal pattern was found in 4.6% of the serum and in 3.6% of the CSF specimens),8 whereas we rarely observed CSF-restricted OCBs in our cohort. This discrepancy is likely ascribed to the relatively recent introduction in laboratory protocols of standardised procedures to identify OCBs.4 Interestingly, only in one patient with CIDP and MGUS IgGk, but with no central nervous system immune-mediated comorbidities, we observed CSF-restricted OCBs in addition to a paraproteinemic pattern. Moreover, CSF-restricted OCBs have also been reported in one patient with axonal neuropathy of the NINPs group, with no other relevant conditions and normal CSF values. We have previously reported CSF OCBs also in two patients with hereditary neuropathy and no other comorbidities23and normal CSF protein concentration and QALB.

Moreover, a substantial percentage of patients with inflammatory neuropathies, in particular patients with GBS (13/32, 40.6%) and CIDP (9/48, 19%) presented a mirror pattern, with systemic OCBs production but absent intrathecal IgG synthesis. These findings are in line with the peripheral aetiology of the disease and suggest that this feature, when considering immune-mediated disorders as a whole, may be more common in CIDP than previously reported.

Potential limits of our study are the retrospective nature and the relative low sample size. However, the homogeneity of the laboratory and CSF analysis performed in the same laboratory and with board certified techniques in consecutive patients from the same centre strengthens the results and the reliability of the data.

In conclusion, based on our population, the presence of a mirror pattern does not seem helpful in differentiating immune-mediated chronic neuropathies from other neuropathies. CSF-restricted OCBs are instead rare in our sample of patients with CIDP (2%), suggesting that their presence might have been overestimated in previous studies, likely due to suboptimal and non-standardised laboratory procedures.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information. Data (deidentified participant data) are available upon reasonable request to the corresponding author.

Ethics statements


CB is part of the European Reference Network for Neuromuscular Diseases. DF is recipient of grant “Progetto Z844A 5×1000” of the Italian Ministry of Health to the IRCCS Ospedale Policlinico San Martino.



  • Contributors MR, MP, MCam, FC and AS had a major role in the acquisition of data, analysed the data and interpreted the data. MP performed the statistical analysis. SR and ET had a major role in the acquisition and interpretation of cerebrospinal fluid (CSF) data. MCac had a major role in the acquisition of neurophysiological data. PG had a major role in the interpretation of CSF data. DF interpreted the data and drafted the manuscript for intellectual content. CB designed and conceptualised the study, coordinated the study, interpreted the data and drafted the manuscript for intellectual content.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles