Neurology India, Vol. 58, No. 4, July-August, 2010, pp. 542-548
Sural nerve biopsy in chronic inflammatory demyelinating polyneuropathy: Are supportive pathologic criteria useful in diagnosis?
Kulkarni GirishB, Mahadevan Anita, Taly ArunB, Nalini A, Shankar SK
Department of Neurology, National Institute of Mental Health & Neurosciences, Bangalore - 560 029
Date of Acceptance: 03-Jun-2010
Code Number: ni10145
AbstractBackground : According to American Academy of Neurology (AAN) criteria, demonstration of demyelination in the sural nerve by teased fiber or ultrastructure is considered mandatory for diagnosis of chronic inflammatory demyelinating polyneuropathies (CIDP). In resource-restricted settings where these techniques are not freely available, it is useful to determine the utility of 'supportive' pathologic criteria (subperineurial edema, inflammation, onion bulb formation, and demyelination) proposed by AAN for diagnosis of CIDP.
Settings and Design : Tertiary care hospital, retrospective study.
Patients and Methods : Forty-six patients with idiopathic CIDP (32 with progressive course and 14 with relapsing-remitting course) satisfying AAN clinical and electrophysiologic criteria evaluated between January 1991 and August 2004 were reviewed. Frequency of specific pathological alterations such as demyelination, inflammation, onion bulb formation, and axonal changes in sural nerve biopsies was evaluated.
Statistical Analysis : SPSS statistical package was used to calculate mean, range, and standard deviation. Student's t test, chi-square test, and ANOVA were used for determining statistical significance.
Results and Conclusion : Reduction in myelinated fiber density was most frequent (93.5%), followed by demyelination (82.8%), inflammation (58.7%), and onion bulb formation (28.3%). Endoneurial inflammation was frequent in the relapsing-remitting form and epineurial inflammation and axonal changes in those with progressive course. Greater disability at presentation, poor response to immunomodulation, and lower CSF protein levels was seen in those with axonal pathology. Pathological abnormalities were demonstrable in all (100%), whereas electrophysiological abnormalities were detected in 90.8%, suggesting that supportive histologic AAN criteria are helpful in diagnosis of CIDP.
Keywords: American Academy of Neurology, chronic inflammatory demyelinating polyneuropathies, diagnosis, neuropathology, supportive pathologic criteria, sural nerve biopsy
Chronic inflammatory demyelinating polyneuropathy (CIDP), first recognized by Austin in 1958,  is clinically a heterogeneous disorder. The classical form is a symmetric, predominantly proximal, demyelinating motor polyneuropathy, but several variants like predominantly distal,  axonal,  or sensory forms,  and asymmetric or focal presentations are described.  Several diagnostic criteria have been proposed based upon clinical and electrophysiological features and cerebrospinal fluid (CSF) and nerve biopsy findings in different combinations ,,, to accommodate the variants. Biopsy evidence of demyelination in sural nerve was considered mandatory in the American Academy of Neurology (AAN) criteria, while subperineurial edema, inflammatory cell infiltration, onion bulb formation, and variation in fascicular involvement were considered supportive of the diagnosis. 
Demonstration of mandatory pathologic criteria requires teased fiber and electron microscopy, both of which are labor intensive and expensive, making their routine use impractical. In resource-restricted settings, it would be useful to evaluate the utility of supportive criteria in the diagnosis of CIDP. We evaluated the contribution of supportive histological features in sural nerve biopsy in ascertaining diagnosis, excluding mimickers, and prognostication in a large cohort of cases clinically diagnosed as CIDP.
Patients and Methods
All patients evaluated between January 1991 and August 2004 and who met the AAN clinical and electrophysiologic criteria for CIDP were included in the study. The clinical findings, demographic details, and treatment response were reviewed by reference to the case records. Patients with retinitis pigmentosa, icthyosis, history of exposure to drugs or environmental toxins known to cause peripheral neuropathy, presence of low serum B 12 , hypothyroidism, family history of neuropathy, as well as cases of amyloid neuropathy, Hansen′s neuritis, and leukodystrophies were excluded.
Investigations essential for establishing the diagnosis of CIDP (CSF examination and electrodiagnostic studies) and for detecting other associated conditions (routine hematology, blood glucose and other biochemical parameters, serological tests for HIV and paraproteinemia, and vasculitis workup) were carried out. The neurological disability was measured according to the Modified Rankin Scale. The modality of therapy used, treatment response, and the clinical outcome were analyzed. All cases recognized to have concurrent illnesses such as diabetes, paraproteinemia, or Polyneuropathy Organomegaly Endocrinopathy M protein, Skin abnormalities (POEMS) were excluded from the study, and only those with idiopathic CIDP (CIDP-I) were analyzed.
Sural nerve biopsies were performed in 46 of the 118 cases clinically diagnosed as CIDP (32 with progressive course and 14 with relapsing-remitting course) after obtaining informed consent. An approximately 1- to 1.5-cm length of sural nerve was biopsied, fixed in 2% gluteraldehyde, and processed for routine paraffin embedding. Sections were stained with hematoxylin and eosin (H and E), Mason′s trichome for collagen, and Kulchitsky Pal (K Pal) stain for myelin. Immunohistochemistry was carried out by indirect immunoperoxidase technique, using antibodies to phosphorylated neurofilament (SMI-31, 1:1000, monoclonal, Sternberger Monoclonal Inc.) to delineate axonal pathology, leukocyte common antigen (LCA) - a leukocyte marker (1:25, polyclonal, DAKO, USA), and CD68 - a macrophage marker (1:200, monoclonal, DAKO, USA).
Biopsies were evaluated for myelin alterations such as active myelin degeneration and presence of onion bulbs (defined as more than one concentric supernumery Schwann cell lamella surrounding the entire circumference of a myelinated axon.  Thin myelin sheaths and reduction in myelinated fiber density on myelin stains was taken as evidence of de/remyelination. Degree of loss of myelinated fibers and variability in fascicular involvement was assessed. Depletion of neurofilament-labeled axons was considered evidence of axonal loss. Endoneurial fibrosis was evaluated on trichrome stain. Axonal regenerating clusters was defined as three or more closely apposed small myelinated fibers with myelin thickness <50% of the thickest fiber in the fascicle.  Findings were considered predominantly demyelinating when axonal changes constituted less than 25% and vice versa. When axonal degeneration and demyelination was 50% ± 10%, the changes were considered mixed, in accordance with previously published studies.  Presence of inflammation and the inflammatory cell type were characterized by immunohistochemistry. Histopathological changes were visually graded as mild (up to 30%), moderate (30%-60%), and severe (>60%), and the results were tabulated.
Mean, range, and standard deviation were derived and significance was calculated using Student′s t test for continuous variables and the chi-square test and ANOVA for categorical variables. Probability level (P-value) <.05 was considered significant.
Among the 118 patients of CIDP diagnosed clinically (male/female ratio 87:31), 91 (77.1%) were idiopathic CIDP (CIDP-I), of whom 46 underwent diagnostic sural nerve biopsy. Age at presentation ranged from 5 to 78 years (mean 37.2 ± 16.7 years). CIDP associated with concurrent disease was identified in 27 (22.9%) patients and were excluded from the study.
Patients with CIDP-I were classified into two major subgroups based on clinical course as progressive (32, 69.5%) and relapsing-remitting (14, 30.4%). Relapse was defined as worsening of symptoms or signs after an initial episode and resulting in an increase in the disability by one or more grades on the disability scale, with subsequent improvement. 
Pathological abnormalities of varying severity were observed in all the nerve biopsies studied [Table - 1].
All four pathological features (subperineurial edema, demyelination, onion bulb formation, and inflammation) were seen in four cases, any three features in 18, two features in 14, and any one feature in ten, though all of them met the clinical and electrophysiological criteria of AAN. Myelin alterations could be demonstrated in all the cases concordant with demyelinating neuropathy, with loss of myelinated fiber density (93.5%) being the most frequent finding followed by evidence of de/remyelination in 82.8% of cases [Figure - 1]b,[Figure - 2]a, and c and onion bulb formation in 28.3% [Figure - 1]c. Axonal loss was evident in in 56.6% of cases [Figure - 2]b and d. Endoneurial Schwann cell proliferation and focal onion bulb formation was more frequent in the progressive than the relapsing-remitting group (33% vs 21%).
Subperineurial edema [Figure - 1]a was noted in more than half the cases, with no significant difference between relapsing-remitting and progressive forms. Inflammation in the endoneurial and epineurial compartment was seen in 58.7% of cases [Figure - 1]d and e, essentially lymphocytic in 50%-60% [Figure - 1]d, inset, and lymphohistiocytic in 10%-18% of cases [Figure - 1]e, inset. Endoneurial inflammation was more evident in the remitting-relapsing form of CIDP-I [Figure1]d, inset in contrast to perivascular epineurial inflammation [Figure - 1]e, inset which was more common in the progressive form of CIDP [Table - 1].
Overall, the disease pattern was predominantly demyelinating in 65.2% (30/46) [Figure - 2]a and b, axonal in 8.7% (4/46) [Figure - 2]c and d, and mixed axonal and demyelinating in 26.1% (12/46). Histopathological abnormalities were demonstrable in all the cases (100%) in contrast to electrophysiological alterations that were observed in 90.8% of cases. Correlation of pathological features with nerve conduction studies showed concordance in 37 patients (68.5%) and discordance in 17 (31.5%). Clinically, sensory symptoms in the lower limbs were observed in 70.4% of patients and sensory signs in 80%, suggesting that pathological alterations precede electrophysiological abnormalities and clinical features.
Endoneurial inflammation and demyelination was more common in the relapsing-remitting form. In contrast, purely axonal changes in biopsy were noted in four cases (8.7%), all of whom had progressive course. Disability rated by Modified Rankin Scale was highest in patients with predominant axonal pathology and this difference was maintained during follow-up visits, with poorer response to immunomodulation, and lower CSF protein levels compared to demyelinating or mixed forms. Mean CSF protein levels were highest (188.9 ± 182.2 mg%) in cases with mixed demyelination and axonal changes on electrophysiology followed by demyelinating (129 ± 97.9 mg%) and axonal group of patients (112.3 ± 73.5 mg%).
Several studies have evaluated the usefulness of the mandatory pathologic AAN criteria ,,,,, in the diagnosis of CIDP. While some believe that sural nerve biopsy is of limited value, , others support the view that nerve biopsy is valuable in atypical presentations  or when electrodiagnostic criteria are not met. , The frequency of detection of various specific pathological features appear to be highly variable in different studies [Table - 2]. Variability in the pathological findings is determined by duration of disease, response to treatment, and the nerve chosen for biopsy. Autopsy studies have demonstrated that inflammation and demyelination more often involves the spinal radicals in a patchy multifocal manner and it may be completely lacking in the segment of distal (e.g., sural) nerve sampled.
The major limitation of the specified mandatory histologic criteria as proposed by the AAN is suboptimal sensitivity and specificity. The key diagnostic neuropathological feature of CIDP is unequivocal demonstration of demyelination that requires teased fiber preparation and electron microscopy. Demyelination on nerve fiber teasing was found in 48%-88% of cases in various studies. ,,, Electron microscopic evidence of demyelination was found in 79%  of cases, including a subset of patients who did not meet the electrodiagnostic criteria for demyelination, suggesting higher sensitivity of nerve biopsies in the diagnosis.  In our study, where we assessed supportive criteria alone, histopathological abnormalities were demonstrable in 100% of sural nerves, compared to demonstration of abnormality in electrodiagnostic features in 90.8%. Inflammatory infiltration of endoneurium by lymphomononuclear cells and macrophages that initiate demyelination , is observed only in a minority (10%-18%) of cases in the sural nerve, ,,, as significant perivascular inflammation is often confined to the proximal nerve roots and sensory ganglia. , In our study, inflammation was demonstrable in a significant number (58.7%) by immunohistochemistry similar to other studies. ,, In the present study, endoneurial inflammation was frequent in the remitting-relapsing form, in contrast to epineurial perivascular inflammation in the progressive form of CIDP-I. It is tempting to speculate that the more prominent axonal pathology, greater disability, and poor response to therapy in the progressive group could be a consequence of epineurial perivascular inflammation producing a sequence of pathogenetic events, similar to what is seen in vasculitic neuropathy.
In the literature, the combination of inflammation with segmental demyelination is reported in less than half the cases. ,, The universal involvement of sural nerves seen in our study is rather surprising as other studies ,, have reported 10%-15% of biopsied nerves to be normal. This may be due to the extensive search we made for pathology, the clinical stage at which biopsy was carried out, or changes coexisting in the nerves that were unrelated to CIDP. Onion bulbs, a characteristic feature in CIDP is reported in 15%-40% of cases. , Dyck and Engelstad  reported a mixed pattern of distribution, with well-developed large onion bulbs intermixed with smaller evolving forms as a characteristic finding in CIDP. In our study, onion bulbs were demonstrable in 28.3% of cases, more commonly in the progressive than in the relapsing-remitting form. Evidence of axonal degeneration and sprouting can accompany de/remyelination ,,, and is considered a bystander effect. Axonal pathology in our study predominated in the progressive form and these patients had greater disability at presentation, poorer response to immunomodulation, and lower CSF protein, as has also been reported by others. , Axonal changes at onset of disease have been found to be a poor prognostic indicator, requiring change to cytotoxic therapy,  and more frequent relapses were noted in this group. 
None of these pathological findings are however specific to CIDP. Endoneurial edema and infiltration by T cells are detected in patients with HMSN type I.  Bosboom et al.  and Molenaar and colleagues  found considerable overlap of findings between CIDP and chronic idiopathic axonal neuropathies (CIAP), including presence of T cells, supporting their view that sural nerve biopsy has limited diagnostic value in CIDP. ,, In recent literature, more specific alterations in nerve biopsies are reported to differentiate CIDP from other inflammatory neuropathies, particularly vasculitis; these include signs of T cell activation, detection of matrix metalloproteinases 2 and 9, ,, chemokine receptors and interferon-γ-inducible protein (IP-10), and upregulation of Th1 cytokine IL-17 and IFN-γ. Macrophage differentiation antigens and ′macrophage clustering′ (defined as presence of three or more macrophages around a blood vessel) around endoneurial vessels have been used to differentiate between inflammatory and hereditary neuropathies. ,
Several authors have reported cases of CIDP that do not fulfill the electrophysiologic criteria of the AAN but who responded to immunomodulatory therapy ,,, and suggest that sural nerve biopsies are indicated in clinically suspected cases of CIDP not satisfying the electrodiagnostic criteria for CIDP. Gabriel and coworkers  documented change in diagnosis following nerve biopsy in 14% of their patients.
Based on available evidence, it is emphasized that while sural nerve biopsy may provide supportive evidence for diagnosing CIDP, the findings are not highly specific and their absence does not exclude a diagnosis. Sensitivity of detection requires deployment of several advanced neuropathological techniques, some of which are not available in many laboratories in India. The European Federation of Neurological Society/Peripheral Nerve Society (EFNS/PNS) recommend the use of semithin sections, electron microscopy, and teased-fiber study for optimum reporting but this is not mandatory.  Despite these drawbacks, nerve biopsy will remain useful to support a diagnosis of CIDP in atypical cases and help exclude other causes of neuropathy like hereditary neuropathies, vasculitis, and infectious neuropathies, and amyloidosis, particularly in cases without family history and where facility for molecular genetic tests are not available. The clinical distinction is crucial because of potential therapeutic implications. Only one other study  that assessed the usefulness of supportive criteria for diagnosis of CIDP concurs with our study, supporting the view that even in the absence of advanced techniques for nerve biopsy processing, detection of supportive features of CIDP still aids in diagnosis, prognostication, and management in developing countries.
The authors wish to acknowledge the technical assistance provided by Mrs. V. Rajyasakti and Mr.Shivaji Rao. The assistance in photographic documentation by Mr. K. Manjunath, Human Brain Tissue Repository (Brain Bank), Department of Neuropathology, National Institute of Mental Health and Neurosciences, Bangalore, is gratefully acknowledged.
Copyright 2010 - Neurology India
The following images related to this document are available:
Photo images[ni10145f2.jpg] [ni10145t2.jpg] [ni10145f1.jpg] [ni10145t1.jpg]