Abstract
Objective: To determine whether microscopic vasculitis explains the clinical and pathologic features of diabetic lumbosacral radiculoplexus neuropathy (DLSRPN).
Background: DLSRPN is usually attributed to metabolic derangement or ischemic injury, but microscopic vasculitis as the sole cause needs consideration.
Methods: We prospectively studied the clinical, laboratory, and EMG features as well as the pathology of distal cutaneous nerve biopsy specimens of patients with DLSRPN.
Results: Study of DLSRPN nerve biopsy specimens (n = 33) compared with those from healthy controls (n = 14) and those with diabetic polyneuropathy (n = 21) provided strong evidence for ischemic injury (axonal degeneration, multifocal fiber loss, focal perineurial necrosis and thickening, injury neuroma, neovascularization, and swollen fibers with accumulated organelles), which we attribute to microscopic vasculitis (epineurial vascular and perivascular inflammation, vessel wall necrosis, and evidence of previous bleeding). Segmental demyelination was significantly associated with multifocal fiber loss.
Conclusions: 1) This severe, debilitating neuropathy begins with symptoms unilaterally and focally in the leg, thigh, or buttock and spreads to involve the other regions of the same and then opposite side and is due to multifocal involvement of lumbosacral roots, plexus, and peripheral nerve (i.e., diabetic lumbosacral radiculoplexus neuropathy). 2) Motor, sensory, and autonomic fibers are all involved. 3) Ischemic injury explains the clinical features and pathologic abnormalities of nerve. 4) The proximate cause of the ischemic injury appears to be microscopic vasculitis. 5) The segmental demyelination is probably secondary to ischemic axonal dystrophy, thus providing a unifying hypothesis for both axonal degeneration and segmental demyelination.
Neuropathy is frequent in diabetes mellitus (DM)1,2 but severe in only a small percentage of patients.2 One severe variety is a subacute, painful, asymmetric lower-limb disorder called “paralytic” or “motor neuropathy,”3,4 “diabetic myelopathy,”5 “diabetic amyotrophy,”6,7 “femoral or femoral-sciatic neuropathy,”8,9 the “Bruns-Garland syndrome,”10,11 “diabetic polyradiculopathy,”12 “diabetic mononeuritis multiplex,”13,14 “proximal diabetic neuropathy,”15,16 and “diabetic lumbosacral plexus neuropathy,”17,18 Opinions vary as to the anatomic distribution and kind of pathologic lesions. Most studies emphasize involvement of the upper lumbar plexus and proximal nerves and of motor fibers. Opinions also vary as to cause: ischemic injury,11,13,14 metabolic derangement,19-22 inflammation,17 necrotizing vasculitis,22-24 or inflammatory demyelination.25 Most authors do not propose a unifying hypothesis and have attributed some cases to ischemia and others to metabolic derangement. Two retrospective studies have reported improvement with immunomodulating therapy,25,26 but efficacy of such therapy remains unproven.
Here we study the clinical features and the pathologic alterations of distal cutaneous nerves of 33 patients with subacutely developing unilateral or asymmetric lower-limb neuropathy not caused by space-occupying lesions in patients with DM, focusing on the underlying pathologic mechanisms of the disorder.
Methods.
Patient selection.
We prospectively studied 33 diabetic patients with static or worsening lumbosacral radiculoplexus neuropathy (DLSRPN) evaluated between April 7, 1995, and January 9, 1998, on whom a distal cutaneous nerve biopsy (sural or superficial peroneal) had been obtained. All patients had DM by National Diabetes Data Group criteria (two fasting plasma glucose values ≥140 mg/dL or had established DM on treatment). Excluded were patients with known causes of plexus neuropathy other than DM, i.e., radiation, infection, bleeding, lymphoma, perineuroma, etc. Although not encountered, symmetric DLSRPN would have been included. Patients were included irrespective of whether the predominant clinical involvement was localized to the buttock, thigh, or leg and was attributed to a DLSRPN. Not excluded were patients who had concurrent median neuropathy at the wrist, ulnar neuropathy at the elbow, cervical radiculoplexus neuropathy, or thoracolumbar radiculoneuropathy.
Neuropathic evaluation.
In addition to a medical and neurologic history and examination, neuropathic impairment was quantitated using the Neuropathy Impairment Score (NIS).27 MRI or CT myelography was used to exclude structural disease of the cauda equina, segmental nerves, or lumbosacral plexus. All patients had characterizing nerve conduction and needle EMG examinations. Neuropathic data were available for three times–at onset (determined from clinical history), at Mayo Clinic Rochester (MCR) evaluation, and at subsequent telephone interview of 31 of 33 patients (January 1998).
Laboratory methods.
We assessed fasting blood glucose, glycated hemoglobin, urea nitrogen, creatinine, rheumatologic tests, and CSF constituents. Also assessed were antibodies to cytomegalovirus (CMV) (CMV IgG, CMV IgM), hepatitis B (HBsAg, anti-HBs, anti-HBc, HBeAg, and anti-HBe), hepatitis C (anti-HCV), Lyme disease, HIV, and HTLV-I; results from 28 of our patients were compared with 28 controls from the Rochester Diabetic Neuropathy Study (RDNS) cohort matched for age, sex, and type of DM.
The RDNS is a prevalence cohort shown to be representative of community diabetics mainly of northern European extraction.28 Cytokine levels (interleukin [IL]-1β, IL-6, and tumor necrosis factor alpha [TNF-α]) were also measured and compared with laboratory controls and our own RDNS controls matched for age, sex, and DM type (IMMULITE IL-1β kit, the IMMULITE IL-6 kit, and the IMMULITE TNF-α kit [EURO/DPC Ltd., Llanberis, Gwynedd, UK]).
Electrophysiologic and quantitative sensory and autonomic testing methods.
Nerve conduction and needle EMG studies were performed using laboratory normal values. The choice of nerves and muscles studied was based on symptoms and findings. Quantitative sensation testing was performed using computer-assisted sensory examination (CASE IV)29,30 on the dorsal foot, lateral leg, or anterior thigh and expressed as percentile values.31 Quantitative autonomic testing was performed using an autonomic reflex screen,32 which measures postganglionic sudomotor, adrenergic, and cardiovagal function.
Histologic methods.
The pathologic alterations of biopsy specimens of distal cutaneous nerves (31 sural and 2 superficial peroneal) were compared with 14 age-matched healthy control sural nerves and with 21 sural nerves from patients with symmetric diabetic polyneuropathy (DPN) matched for age and DM type. In 32 nerves, paraffin sections were stained, using methods usual in our laboratory, for leukocytes (leukocyte common antigen) and macrophages (KP-1). One nerve biopsy specimen was obtained from an outside institution and reviewed. Semi-thin epoxy sections were stained with methylene blue and p-phenylenediamine. Electron microscopy was performed on unrecognized structural profiles. Teased fibers from control and disease nerves were graded (n > 100 myelinated fibers or breakdown products per nerve) with the identification of nerve slides masked from the observer (P.J.B.D.) by previously defined pathologic alteration criteria.33 In fixed sections, the pathologic alterations were graded semi-quantitatively for features of ischemic injury and vasculitis.
Analysis.
Descriptive statistics were used to express results and to compare attributes between groups. For continuous measurements we expressed results as medians and ranges and compared groups using Wilcoxon rank sum tests. For dichotomous variables we used Fisher’s exact test.
Results.
Diabetic and laboratory characteristics.
Our cohort had a median age of 65 years (range 36 to 76), had DM for a median time of 4.1 years (range 0 to 36), and had a median glycated hemoglobin of 7.5% (range 5.1 to 12.9%). Only one patient had type 1 DM but 13 used insulin. Of this group, only four had retinopathy (nonproliferative), and two had nephropathy. Although the group still tended to be overweight (median body mass index 25.7, range 17.8 to 36.7), their median weight loss was 30 pounds (range 0 to 120). Other clinical and laboratory features are listed in table E-1, website. The DLSRPN cohort, when compared with the RDNS cohort, was not significantly different in age, but more had type 2 DM (p = 0.005). Their glycemic control was significantly better (glycosylated hemoglobin was 2.3% lower) (p = 0.001); their body mass index was significantly lower (by ∼3.0) (p = 0.002); and they had a higher male-to-female ratio. Weight loss (≥10 pounds) was recorded for 28 of 33 patients and was significantly more common than it was in the RDNS cohort (p = 0.001). They had significantly less retinopathy (p = 0.001) and cardiovascular disease (p = 0.001).
Of the rheumatologic, serologic, and antibody tests performed, no consistent pattern of abnormalities was found (see table E-1, website). Typically, the CSF had a markedly increased protein concentration (median 89 g/dL, range 44.0 to 214.0) but a normal number of cells. Cytokine levels (IL-1β, IL-6, and TNF-α) were increased above laboratory normal values, but when compared with RDNS patients, only IL-1β and IL-6 approached statistical significance (see table E-1, website).
Characterization of the neuropathy.
The characteristic symptoms were asymmetric lower-limb pain (hurting, tightness, lancinating, burning, and allodynia) (all 33), weakness and atrophy (all 33), and paresthesia (31 of 33) (table 1). At onset, pain was listed as the most common and most severe symptom. By the time of MCR evaluation (median 6.7 months from onset, range 1.4 to 42.0), weakness had become the most prominent symptom. Weakness continued to be the most bothersome symptom at recent telephone interview (median 25.9 months from onset, range 4.5 to 46.5). One-half the patients (n = 16) had symptoms of autonomic dysfunction, which included orthostatic hypotension, urinary dysfunction, constipation, diarrhea, tachycardia, and new-onset impotence.
Severity of symptoms, anatomic location, and disability in patients with DLSRPN at three times
The most severe symptom (pain, weakness, or paresthesia) initially and at time of MCR evaluation was more often in the thigh than in the leg or foot and was rarely in the buttock or back. However, at telephone follow-up, the most severe symptoms had shifted to the leg from the thigh (see table 1). Although symptoms usually began either focally in the thigh or leg, they almost always became more generalized to involve the entire lower limb and then spread to involve the contralateral limb. Thus, at onset, 29 patients had unilateral pain, weakness, or paresthesia; however, by the time of our evaluation, all but 1 had bilateral symptoms and findings (see table 1). The median time to bilateral involvement was 3.0 months (range 0 to 60 months). The median NIS (43.0 points, range 7.0 to 87.0) (table E-2, website) indicates that the neuropathy was severe and was much worse than the median NIS of DPN patients in the RDNS cohort (2.0 points, range 0 to 18.0) (p = 0.001).
Our patients experienced considerable morbidity; all sought medical help, essentially all received analgesic medication, and 25 patients required narcotic medication. At MCR evaluation, about one-half required the use of wheelchairs (n = 16), and most others needed a walker, cane, or leg brace (n = 14). At telephone follow-up, only 2 of 31 patients reported that they had returned to normal. Most patients still had troublesome degrees of pain and weakness. Nevertheless, real improvement had occurred; only 3 still used wheelchairs, 16 used aids for walking, and 12 required no aids.
Four patients had additional symptoms and findings consistent with the diagnosis of thoracolumbar radiculopathy, with a band of pain around their chest or abdominal wall and weakness. One third (n = 11) of the patients also had problematic neuropathic symptoms and findings in their upper limbs. Based on clinical history, neurologic examination, and electrophysiologic testing, 8 patients had upper-limb mononeuropathies (6 focal ulnar neuropathy at the elbow [3 unilateral and 3 bilateral] and 2 median neuropathy at the wrist). Three patients had bilateral but asymmetric cervical radiculoplexus neuropathies.
Electrophysiologic findings.
The nerve conduction abnormalities in lower limbs consisted of reduction or absence of the compound muscle action potential (CMAP) of tibial and peroneal nerves and of the sensory nerve action potential (SNAP) of the sural nerve with only a mild reduction in conduction velocities (see table E-2, website). A characteristic feature was the asymmetry of CMAP and SNAP amplitudes between sides. From the distribution of fibrillation potentials and motor unit potential changes on needle EMG examination (see table E-2, website), involvement was generally more widespread than had been recognized from the clinical examination and was in keeping with an asymmetric and multifocal pathologic process in lumbar or lumbosacral roots, segmental nerves, plexus or lower-limb nerves. The changes were not confined to the territory of one peripheral nerve (i.e., femoral). From the nerve conduction and EMG abnormalities, we infer that axonal degeneration is more severe than segmental demyelination.
Quantitative sensory testing.
Seventeen patients were assessed for vibration, cooling, and heat pain thresholds at three anatomic sites of the lower limb by CASE IV.29,30 Results are expressed as low (hyperesthesia or hyperalgesia), normal, or high (hypoesthesia or hypoalgesia) thresholds (table 2).31 Hyperalgesia was found in three of 26 heat pain tests, whereas hypoalgesia was found in 10 of 26 tests. For vibration and cooling detection thresholds, none of the values were low (hypersensitivity). For vibration, 19 of 27 patients had raised thresholds (hyposensitivity). For cooling, 13 of 22 also had raised thresholds. These results show that there is an unequivocal sensory abnormality at different anatomic sites to all sensory modalities.
Quantitative sensory and autonomic test results in DLSRPN patients at Mayo Clinic evaluation
Quantitative autonomic testing.
Of 14 patients tested, 8 had clinical symptoms of autonomic dysfunction. A composite autonomic severity score (CASS)32 was mildly abnormal in 4, moderately abnormal in 2, and severely abnormal in 8 of 14 patients. The median CASS (see table 2) showed a moderate to severe autonomic dysfunction overall, whereas the median sudomotor, adrenergic, and cardiovascular indices were all increased, demonstrating that the autonomic dysfunction was generalized. The four patients who had thermoregulatory sweat distribution tests all showed patchy anhidrosis of affected lower limbs.
Pathologic alteration of distal cutaneous sensory nerves.
The pathologic evidence for nerve ischemia (table 3) consisted of 1) focal or multifocal fiber degeneration or loss in 19 nerves (figure 1); 2) focal degeneration of the perineurium in 6 nerves or focal thickening or scarring of the perineurium in 24 nerves; 3) epineurial neovascularization in 21 nerves; and 4) abortive regeneration of nerve fibers outside of the original perineurium, forming microfasciculi (injury neuroma) in 12 nerves (see figure 1). The injury neuromas and perineurial scarring often were associated with damaged parent fascicles showing focal fiber loss. Occasional fibers showed enlarged dark axons (caused by accumulation of organelles) that contained light cores (remaining axoplasm). Sometimes these “dark axons with light cores” had become demyelinated (figure 2). These changes were significantly more frequent and severe in DLSRPN nerves than they were in DPN or healthy control nerves (see table 3).
Pathologic results of distal cutaneous nerve biopsy specimens from patients with diabetic lumbosacral radiculoplexus neuropathy (DLSRPN) compared with diabetic polyneuropathy (DPN) and healthy controls
Figure 1. Transverse epoxy sections (p-phenylenediamine) of distal sural nerves from patients with diabetic lumbosacral radiculoplexus neuropathy illustrating the dramatic focal fiber loss characteristic of the disorder (A, fascicle on the left) and the abortive microfascicular nerve regeneration (B, as identified by the arrow). Note that the abortive regeneration (injury neuroma) is made up of multiple regenerating fascicles and that they are situated adjacent to a fascicle devoid of myelinated fibers. Most of the fibers in the right fascicle in the lower panel are actively degenerating. As discussed in the text, these changes are indicative of ischemic injury that we attribute to a microscopic vasculitis.
Figure 2. Transverse electron micrograph of an enlarged “dark axon with light cores” from the sural nerve of a patient with diabetic lumbosacral radiculoplexus neuropathy. These changes are typical of pathologic alterations of nerve fibers near the proximal and outer edge of ischemic cores in animal models of ischemia. The axon is enlarged and demyelinated and is full of accumulated organelles. The “light core” (asterisk) is a preserved area of axoplasm. As discussed in the text, we hypothesize that ischemia causes an interruption of fast axonal transport, accumulation of organelles, axonal enlargement, and local demyelination. Depending on severity, the distal axon may survive or may degenerate.
Epineurial perivascular and vascular mononuclear inflammatory cell collections were found in all nerves and were significantly more frequent than they were in the DPN or healthy control nerves (see table 3). The mononuclear cells reacted for leukocyte common antigen in all (32 of 32) nerves and to a lesser extent for a macrophage marker (KP-1). Features diagnostic of necrotizing vasculitis (necrosis of the vessel wall associated with vascular and perivascular inflammatory cells) were found in two nerves, and suggestive changes of necrotizing vasculitis (inflammatory cells infiltrating vessel walls) (figure 3) were seen in an additional 13 nerves compared with none in the healthy or DPN control nerves. The involved vessels were small arterioles, venules, and capillaries. Hemosiderin (evidence of previous microscopic hemorrhage) was found in 19 of 32 nerves (Turnbull blue stain) (see table 3) but in none of the two control groups. Frequently (n = 14), the hemosiderin was in the sub-perineurium or in the perineurium, often adjacent to microvessels and was distributed multifocally. Often, the hemosiderin was associated with inflammation and sometimes with scarred epineurial arterioles (see figure 3).
Figure 3. Transverse paraffin sections of sural nerves from patients with diabetic lumbosacral radiculoplexus neuropathy (DLSRPN) showing epineurial microscopic vasculitis. (A) (Hematoxylin-eosin) Mononuclear cell infiltration and separation of smooth muscle of a small epineurial arteriole. The arrow shows a deposit of hemosiderin. (B) (Turnbull blue) An epineurial arteriole with intimal thickening, adventitial scarring, recanalization, and perivascular hemosiderin deposition. The asterisk shows fresh blood (pink), whereas the arrow shows previous bleeding (hemosiderin) that stained blue. The arrowhead shows intimal proliferation. As described in the text, vascular and perivascular epineurial inflammation of small arterioles and hemosiderin occurred in most nerves. When taken with the ischemic pathologic changes, these findings provide strong evidence that microscopic vasculitis is the responsible underlying mechanism in DLSRPN.
Axonal degeneration and empty nerve strands were the most frequent teased fiber abnormalities seen (see table 3). A mildly increased rate of demyelination was also observed. Often, areas of demyelination and remyelination were clustered on individual teased fibers, suggesting secondary demyelination to underlying axonal pathology (figure 4). Of the 12 biopsy specimens that showed an increased rate of demyelination (>5%), 11 also had evidence of multifocal fiber loss (p = 0.01).
Figure 4. Eleven contiguous regions along the length of a single teased myelinated fiber from the sural nerve of a patient with diabetic lumbosacral radiculoplexus neuropathy showing multiple areas of demyelination and remyelination. Segmental demyelination is seen between arrows in B, C, E through F (note myelin ovoid in demyelinated nerve strand in E), G, and K, and remyelination (areas of myelin that are less than 50% of the thickness of the thickest myelin segments) in A, B, C, D, and K. This clustering of demyelination and remyelination along the same nerve fiber was seen frequently in nerves of DLSRPN and suggests that the demyelination is secondary to axonal dystrophy (as described in the text).
Discussion.
Although most previous investigators separate the subacute, asymmetric, painful, lower-limb diabetic neuropathy (DLSRPN) from the symmetric distal DPN, there is still some uncertainty about the clinical features and course, the outcome, the classes of fibers affected, the proximal to distal level of nerve involvement, and the putative pathologic process responsible. Most authors write that DLSRPN is part of a continuum with rapid onset and asymmetric features (perhaps from ischemia) at one extreme and a slowly evolving, symmetric disorder (perhaps from metabolic derangement) at the other extreme.16,21,34
We found, as had previous investigators, that the disorder typically affects older people with mild type 2 DM and is associated with weight loss. The major symptoms are pain and weakness, and lesser symptoms are paresthesia, allodynia, autonomic dysfunction, and muscle atrophy. Typically the disorder begins with pain, but after a short time, weakness predominates over pain. We emphasize that the disorder characteristically has a definite date of onset, begins focally and unilaterally or asymmetrically (all cases), and affects the leg almost as frequently as the thigh but then usually progresses to involve the other segment of the lower limb and then spreads to the contralateral limb. Consequently the terms proximal or femoral neuropathy seem inappropriate. The frequent occurrence of distal and bilateral involvement has been emphasized by others.11,34 Some of the differences in reports of unilateral versus bilateral involvement could be explained by what point in their disease course patients were evaluated. As we show, if seen early, the condition might be unilateral; if seen late, it might be bilateral.
Also, unlike what was thought by many earlier workers,7,14,19,24 this disorder is not primarily a motor neuropathy; as we show here, sensory fibers of all modalities and autonomic fibers of all systems are unequivocally affected. Most of the sensory involvement cannot be attributed to a concomitant distal DPN because symptoms begin with the onset of DLSRPN, and its distribution is not length dependent.
Perhaps insufficiently emphasized in previous reports are the long duration and severity of the pain and weakness and the ensuing disability. Pain was often persistent and severe, interfering with work and recreation and necessitating hospitalization. Most patients used narcotic medication, and one half still had a problematic degree of pain 2 years after the onset of the disorder. Weakness was a minor symptom at onset, but by the time of our evaluation it was the major problem and recovery was slow and incomplete. However, by the time of our recent telephone interview, real improvement had occurred, and few patients were still in wheelchairs. The pattern of the weakness changed over time. Initially, proximal limb weakness predominated, but later distal weakness predominated. This finding probably indicates that reinnervation occurs earlier and more effectively in proximal than in distal muscles.
Based on clinical features, EMG, elevated CSF protein, and distal cutaneous nerve biopsy results, we infer that the responsible pathologic abnormalities are multifocal and in roots, segmental nerves, lumbar or lumbosacral plexus, and (as we show here) in peripheral nerves (i.e., a DLSRPN).
In a minority of cases, thoracolumbar radiculopathies and upper-limb neuropathies accompanied the DLSRPN. Most of the upper-limb involvement was due to entrapment or compression of the median nerve at the wrist or the ulnar nerve at the elbow. However, there were three cases of bilateral, asymmetric cervical radiculoplexus neuropathies analogous to the lower-limb syndrome.
We conclude from our studies that the main pathologic process is ischemia from microscopic vasculitis and that this process explains both the axonal degeneration and segmental demyelination observed. In the biopsy specimens of distal nerves, we found a constellation of pathologic findings, usually localized to the same region of the nerve, indicative of ischemic injury. Multifocal fiber loss has been frequently described in this condition and is attributed to ischemia.11,13,14,17,22,23 Other evidence of ischemic injury, which we found, includes the following: axonal degeneration, focal necrosis or scarring (thickening) of the perineurium, abortive nerve regeneration (injury neuroma) often disrupting the perineurium, new blood vessel formation, and enlarged dark axons with accumulated organelles with light cores (preserved axoplasma). Although these abnormalities occur frequently, they have either not previously been reported or have only been reported rarely (injury neuroma in one case).25 But these pathologic findings have been described in necrotizing vasculitis in humans35 and in animal models of ischemia36-41 and are not characteristic of inflammatory-demyelinating or metabolic processes.
The thinking about the role of inflammation and specifically the role of microscopic vasculitis in DLSRPN has evolved. Early workers demonstrated inflammation13,14 but downplayed its importance, whereas others17 deliberately excluded cases with necrotizing vasculitis but reported perivascular inflammatory cell cuffing. Recent authors22,24 attribute some cases to necrotizing vasculitis but still believe other cases are due to metabolic mechanisms. Others42 report increased inflammatory cells (mainly CD8) in both DPN and DLSRPN.
We find considerable evidence that the primary event in DLSRPN is a microscopic vasculitis: epineurial perivascular mononuclear inflammation (found in all nerve biopsy specimens); vessel wall inflammation (found in approximately one-half of nerves); previous bleeding (found in more than one-half of nerves); and changes diagnostic (n = 2) or suggestive (n = 13) of necrotizing vasculitis. These changes were not found in DPN or in healthy control nerves. The pathologic alterations seen here involve small arterioles, venules, or capillaries (i.e., a microscopic vasculitis). We only rarely encountered the florid features of necrotizing vasculitis of large arterioles and arteries (e.g., necrosis of tunica media with inflammation) typical of other systemic vasculitides. Because the affected vessels in DLSRPN have thin walls, fibrinoid necrosis of the vessel is not usually seen; only a punctate hemorrhage (hemosiderin) or focal inflammation remains. Microscopic vasculitis is known to present with these findings in other tissues.43 Finding hemosiderin in most cases strongly favors a vasculitic over an inflammatory demyelinating or metabolic cause. We also suggest that the weight loss that is so typical of this disorder5-7 may be a constitutional symptom of necrotizing vasculitis and is unlikely to be due solely to poor metabolic control, deliberate weight loss, depression, pain, or use of analgesics or narcotic medications.
Previous investigators have attributed axonal degeneration to ischemia and segmental demyelination to a metabolic disturbance. We propose, in a unifying hypothesis, that both the axonal degeneration and the segmental demyelination, as seen here, are the result of ischemic injury from microscopic vasculitis. We base this hypothesis on the following five lines of evidence: 1) we did not find two subtypes of DLSRPN, one ischemic (axonal) and the other metabolic (demyelinating); 2) segmental demyelination was statistically significantly linked to regions of ischemic injury (focal fiber loss); 3) the level of hyperglycemia was not sufficient to explain demyelination (assuming that hyperglycemia causes demyelination); 4) the clustered pattern of demyelination we found is characteristic of demyelination secondary to axonal dystrophy as we have reported in certain human neuropathies44,45 and animal models46; and 5) the axonal enlargements and accumulated organelles typical of ischemic injury,37 which we found, would explain demyelination from swelling or atrophy of axons. The presence of axonal degeneration and segmental demyelination together in the same region is typical of the pathologic changes at the edges of ischemic cores in animal models.36,37 These changes have been attributed to interference of rapid axonal transport due to ischemia. The observation in DLSRPN, which we and others23 have made of proximal demyelination and distal axonal degeneration on the same teased fiber, is direct evidence that the two processes (demyelination and axonal degeneration) are linked. We infer that the clinical course, anatomic distribution of the neurologic and EMG findings, and the pathologic changes are best explained by a unitary process—ischemic injury from microscopic vasculitis that results in both axonal degeneration and segmental demyelination.
Acknowledgments
Supported in part by a grant from NINDS (36797).
Acknowledgment
The authors thank Mary Ann Armbruester, PhD, Diagnostic Products, Inc., Los Angeles, CA, and George Klee, PhD, Mayo Clinic Rochester, for the measurements of plasma cytokines; Phillip Low, MD, Mayo Clinic Rochester, for quantitative autonomic test assessment; and Mary Lou Hunziker and Carol Overland, Mayo Clinic Rochester, for preparation of the manuscript. They also thank their patients for their cooperation.
Footnotes
-
Presented at the 123rd annual meeting of the American Neurological Association; Montreal, Canada; October 1998.
- Received September 29, 1998.
- Accepted August 18, 1999.
References
Disputes & Debates: Rapid online correspondence
REQUIREMENTS
If you are uploading a letter concerning an article:
You must have updated your disclosures within six months: http://submit.neurology.org
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.