Introduction
Many varieties of neuropathy occur with diabetes mellitus (DM) from different underlying mechanisms, the most common being a symmetrical, length-dependent, sensorimotor polyneuropathy (DSPN). A less frequent diabetic neuropathy which can cause severe morbidity, diabetic lumbosacral radiculoplexus neuropathy (DLRPN), is a painful, rapidly evolving, asymmetric, lower-limb, paralytic neuropathy associated with weight loss1. The name DLRPN emphasizes the association of this syndrome with DM and its sites of involvement (roots, plexus and nerves).
The natural history and underlying mechanisms of DLRPN have been debated with various names reflecting different views that include: diabetic myelopathy2, diabetic amyotrophy3, Bruns-Garland syndrome4, diabetic mononeuritis multiplex5, proximal diabetic neuropathy6, diabetic lumbosacral plexopathy7, diabetic polyradiculopathy8 and multifocal diabetic neuropathy9. Regardless, in most discussions of DLRPN, pain is usually present.
In early clinical descriptions of DLRPN, Chokroverty and colleagues state that this syndrome can be symmetrical, slowly progressive and occasional painless10. They and others suggest that the pathological basis of DLPRN is probably metabolic derangement10–13. Others note that the disease is usually asymmetrical and painful, presenting as a “mononeuritis multiplex” due to ischemic injury5. Asbury incorporates both concepts proposing that the acute and asymmetrical cases are due to ischemic injury whereas the slower, more symmetrical cases result from metabolic derangement and that “proximal (motor) diabetic neuropathy” encompasses a spectrum of disorders6. Additional authors note that cases of purely proximal diabetic motor neuropathies exist, but are poorly defined14.
Recent studies report ischemic injury from microvasculitis as the pathophysiological basis of typical, painful, DLPRN1, 15–19. In contrast, diabetic patients with lower limb motor syndromes without pain have received little attention. Amato and Barohn highlight that motor polyradiculoneuropathies without pain do occur in DM20. Although they consider this painless syndrome could be a type of DLRPN, they question whether such cases are chronic inflammatory demyelinating polyradiculoneuropathy (CIDP) occurring in diabetic patients, an association also suggested by others21–24. Their discussion rekindles interest into how to best classify diabetic patients with a motor predominant painless neuropathy. However, there has been no systematic evaluation of the clinical, electrophysiological, and pathological features of this painless syndrome.
Here, we study a cohort of diabetic patients with a lower limb and motor predominant neuropathy without pain and compare this group to a cohort of previously studied DLRPN patients1. In order to formally answer if painless diabetic motor neuropathy is a variant of DLPRN, diabetic CIDP, or another entity, we ask 1) What are the clinical features of this painless syndrome? 2) What are the laboratory and electrophysiological findings? 3) What are the pathological features? As we believe the classification should be primarily based on the underlying disease mechanism, we place the most emphasis on pathological findings in this group compared to nerve biopsies from DLRPN, DSPN and CIDP.
MATERIAL AND METHODS
Patient selection
After IRB approval, we reviewed medical records of patients seen at Mayo Clinic Rochester in Neurology between January 1, 1985 and December 31, 2005 with coded diagnoses of polyradiculopathy, polyradiculoneuropathy, lumbosacral plexopathy, lumbosacral radiculoplexus neuropathy, proximal diabetic neuropathy, diabetic amyotrophy, or diabetic mononeuritis multiplex from whom a distal cutaneous nerve biopsy was obtained. Neuropathic impairment was quantitated using the Neuropathy Impairment Score (NIS), a global score of weakness, reflex and sensory loss25. 470 patients were identified who were further reviewed for the following clinical and electrophysiological features: 1) documented absence of neurogenic pain or use of analgesic drugs for the disease, 2) weakness in one or both lower limbs, 3) presence of sensory symptoms or signs, 4) nerve conduction studies (NCS) or needle electromyography (EMG) abnormalities involving at least two peripheral nerves from at least two nerve root levels (a lumbosacral plexopathy). Patients were selected irrespective of whether the NCS showed axonal or demyelinating features. Patients could have upper limb or thoracic involvement but the prominent abnormalities were of lower limbs. All patients had DM (two fasting plasma glucose values ≥126 mg/dL) or a prior diagnosis of DM. In all but seven patients, information regarding weight change in relation to the onset of their neuropathic symptoms was documented. Patients who had structural lesions on imaging, other clear causes of neuropathy, or central nervous system disease that interfered with neuropathy evaluation were excluded. Between January 1, 2006 and March 31, 2009 we prospectively identified four additional patients who fulfilled our clinical criteria. We compared the pathological findings in this painless cohort to biopsies from 33 previously reported DLRPN patients1, 21 previously reported DSPN patients1, and 25 CIDP patients who met both our26 and the AAN AIDS Task Force clinical and electrophysiological criteria for CIDP27.
Electrophysiologic, quantitative sensory and autonomic testing
All patients underwent NCS and EMG testing using laboratory normal values. Nine patients had quantitative sensation testing performed using CASE IV28–31 and quantitative autonomic testing, a measure of postganglionic sudomotor, adrenergic, and cardiovagal function32.
Histologic methods
All distal cutaneous nerve biopsies were re-graded by two authors (PJBD, MGS). On all nerve specimens teased fiber preparations; paraffin sections stained with hematoxylin and eosin (H&E); Luxol Fast Blue/Periodic-Acid-Schiff (LFB/PAS), Mason’s trichrome, Congo-red, methyl violet and Turnbull blue (iron stain); epoxy sections stained with methylene blue; and paraffin sections immunostained with CD45, CD68, and Dako anti-human smooth muscle actin (smactin) were reviewed.
Statistics
Descriptive statistics were used to compare attributes between groups. For continuous measurements results were expressed as medians/ranges using Wilcoxon rank sum tests. For dichotomous variables Fisher’s exact test was used.
RESULTS
Demographic and laboratory characteristics
We identified 23 patients who fulfilled our criteria for painless diabetic motor predominant neuropathy. 12 were men and 11 women, with a median age of 62.2 years (range 36 to 78) with DM for a median duration of 5.4 years (range 0 to 23). In seven patients, DM was identified at presentation of neurological illness. All patients except one had type 2 DM; eight were insulin requiring and the rest used oral hypoglycemic medication; all had good glucose control. Chronic diabetic complications (retinopathy, nephropathy) were identified in four patients. Weight loss (≥10 pounds) was similar in severity and frequency between the painless diabetic cohort [13 of 16 (81%)] and the DLRPN cohort [28 of 33 (84%)], however body mass index was significantly higher in the painless group (Table 1). CSF testing performed in 16 patients demonstrated elevated protein with normal cell count. Laboratory results suggestive of an inflammatory process (rheumatoid factor, antinuclear antibodies, erythrocyte sedimentation rate, etc.) were present in a minority of patients. When compared to the DLRPN cohort, the painless group was not significantly different with regards to gender, age, type of DM, glycemic control or associated DM complications (Table 1).
Table 1.
Clinical and Laboratory Characteristics in Painless Diabetic Motor Neuropathy and Painful DLRPN
| Painless Syndrome | Painful DLRPN | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| N | Median | Range | SD | N | Median | Range | SD | p* | |||||
| Clinical | |||||||||||||
| Age, yr | 23 | 62.2 | 36.3 | – | 78.4 | 11.1 | 33 | 65.4 | 35.8 | – | 75.9 | 10.4 | NS |
| Body mass index, kg/m2 | 23 | 29.1 | 18.0 | – | 44.5 | 7.4 | 29 | 25.7 | 17.8 | – | 36.7 | 4.9 | 0.01 |
| Weight change, lb | 16 | −30.0 | −100.0 | – | 0.0 | 29.7 | 33 | −30.0 | −120.0 | – | 0.0 | 32.6 | NS |
| Duration of neuropathy at evaluation, mo | 23 | 11.0 | 2.0 | – | 108.0 | 25.0 | 33 | 6.7 | 1.4 | – | 42.0 | 8.9 | NS |
| Time to bilateral, mo | 21 | 0.0 | 0.0 | – | 36.0 | 7.9 | 32 | 3.0 | 0.0 | – | 60.0 | 10.6 | 0.001 |
| Time to maximum disability, mo | 23 | 10.0 | 3.0 | – | 108.0 | 23.3 | 31 | 5.0 | 0.3 | – | 20.0 | 4.7 | 0.005 |
| NIS Total | 23 | 70.3 | 11.0 | – | 124.0 | 26.6 | 33 | 43.0 | 7.0 | – | 87.0 | 18.6 | <0.0001 |
| NIS Upper limb | 23 | 13.0 | 0.0 | – | 56.3 | 20.4 | 33 | 4.0 | 0.0 | – | 25.5 | 8.7 | 0.005 |
| NIS Lower limb | 23 | 50.0 | 8.0 | – | 70.0 | 13.4 | 33 | 37.0 | 7.0 | – | 62.0 | 14.4 | 0.001 |
| NIS Hip and thigh | 23 | 18.0 | 0.0 | – | 31.0 | 8.5 | 33 | 14.5 | 0.5 | – | 29.0 | 8.1 | NS |
| NIS Leg | 23 | 34.0 | 8.0 | – | 46.0 | 8.4 | 33 | 22.5 | 4.0 | – | 40.0 | 9.6 | 0.001 |
| Laboratory | |||||||||||||
| Fasting plasma glucose, mg/dL | 23 | 150.0 | 84.0 | – | 293.0 | 57.6 | 30 | 144.5 | 75.0 | – | 225.0 | 44.3 | NS |
| Glycated hemoglobin, % | 22 | 7.1 | 4.8 | – | 14.8 | 2.5 | 30 | 7.5 | 5.1 | – | 12.9 | 2.0 | NS |
| Creatinine, mg/dL | 21 | 0.9 | 0.4 | – | 7.4 | 1.5 | 30 | 0.9 | 0.7 | – | 3.4 | 0.5 | NS |
| Cerebrospinal fluid glucose, mg/dL | 23 | 88.0 | 45.0 | – | 156.0 | 28.3 | 26 | 85.0 | 56.0 | – | 130.0 | 19.4 | NS |
| Cerebrospinal fluid protein, mg/dL | 23 | 89.0 | 30.0 | – | 211.0 | 44.3 | 26 | 89.5 | 44.0 | – | 214.0 | 35.3 | NS |
| Cerebrospinal fluid cells, cells/µL | 23 | 2.0 | 1.0 | – | 17.0 | 3.4 | 26 | 1.0 | 1.0 | – | 11.5 | 2.1 | 0.02 |
| Erythrocyte sedimentation rate, mm/h | 19 | 17.0 | 0.0 | – | 53.0 | 13.8 | 31 | 6.0 | 0.0 | – | 60.5 | 14.6 | 0.02 |
| N | Yes | No | N | Yes | No | p | |||||||
| Clinical | |||||||||||||
| Gender, male | 23 | 12 | 11 | 33 | 20 | 13 | NS | ||||||
| Diabetes, type 2 | 23 | 22 | 1 | 33 | 32 | 1 | NS | ||||||
| Insulin use | 23 | 8 | 15 | 31 | 13 | 18 | NS | ||||||
| Retinopathy | 17 | 4 | 13 | 17 | 4† | 13 | NS | ||||||
| Nephropathy | 23 | 3 | 20 | 33 | 2 | 31 | NS | ||||||
| Cardiovascular disease | 22 | 7 | 15 | 33 | 3 | 30 | NS | ||||||
| Laboratory | |||||||||||||
| Rheumatoid factor, reactive | 11 | 2 | 9 | 25 | 1 | 24 | NS | ||||||
| ANA, positive | 18 | 4 | 14 | 32 | 7 | 25 | NS | ||||||
DLRPN = diabetic lumbosacral radiculoplexus neuropathy; SD=standard deviation; NS = not significant (p > 0.05); NIS = Neuropathy Impairment Score.
Wilcoxon rank sum test for continuous data and Fisher's exact test for dichotomous data.
Nonproliferative retinopathy.
Neuropathy Characteristics
The most common clinical pattern seen in this painless cohort was a motor predominant neuropathy developing over weeks to months with bilateral foot drop that progressed to involve proximal lower limb muscles with lesser upper limb involvement (Table 2). The impairment was severe, with all patients experiencing weakness and half of the patients being wheelchair-bound at presentation. 22 of 23 patients had sensory symptoms, 15 with numbness and seven having prickling paresthesias involving the distal lower limb. Nine patients had symptoms of autonomic dysfunction: orthostatic hypotension (2), urinary dysfunction (1), constipation (3), gastroparesis (2), impotence (5) and sweat dysfunction (1). Two-thirds of the patients had bilateral onset of symptoms and almost all had bilateral involvement by the time of evaluation (Table 2).
Table 2.
Neuropathy Characteristics in Painless Diabetic Motor Neuropathy and Painful DLRPN
| Painless Syndrome | Painful DLRPN | ||
|---|---|---|---|
| N = 23 | N = 33 | p* | |
| Onset | |||
| Acute | 0 | 10 | 0.002 |
| Subacute | 21 | 23 | |
| Chronic | 2 | 0 | |
| Onset Pattern | |||
| Bilateral | 15 | 4 | 0.0001 |
| Unilateral | 8 | 28 | |
| Progression | |||
| Rapid (< 6 wks) | 0 | 1 | 0.02 |
| Intermediate (6 – 26 wks) | 8 | 21 | |
| Slow (> 26 wks) | 15 | 9 | |
| Course | |||
| Monophasic | 22 | 25 | NS |
| Polyphasic | 1 | 8 | |
| Most involved site | |||
| Distal (foot or leg) | 13 | 14 | <0.001 |
| Proximal (hip or thigh) | 4 | 19 | |
| Equal (distal and proximal) | 6 | 0 | |
| Buttock or Back | 0 | 0 | |
| Aids in ambulation | |||
| Wheelchair | 11 | 16 | NS |
| Walker, Cane or Brace | 11 | 14 | |
| None | 0 | 3 | |
| Unknown | 1 | 0 | |
DLRPN = diabetic lumbosacral radiculoplexus neuropathy; NS = not significant (p > 0.05).
Fisher's exact test.
Neurological examination showed a largely symmetrical polyradicular pattern with distal lower limb segments being most affected in contrast to DLRPN patients in whom the proximal lower limb is the most severely affected. However, proximal segments were involved to the same degree as DLRPN (Table 1) which is evidence against this painless syndrome simply being a length-dependent process. There was prominent upper limb involvement in 17/23 painless patients that was of a greater severity (higher upper limb NIS) than in DLRPN patients resulting in a significantly higher median NIS in the painless group (Table 1). These findings taken together suggest that the deficits are more severe, more often bilateral at onset and overall more widespread in the painless diabetic motor syndrome as compared to DLRPN.
Physiological Testing
NCS demonstrated reduced/absent upper and lower limb sensory nerve and compound muscle action potential amplitudes without temporal dispersion or conduction block with mild slowing of conduction velocities and mild prolongation of distal and F-wave latencies (Table 3). Needle examination demonstrated patchy but widespread involvement of multiple lumbosacral myotomes with denervation and chronic motor unit potential changes. Lumbosacral paraspinal fibrillation potentials were seen in 14 of the 19 patients tested. Of the nine patients who had quantitative sensory and autonomic studies performed, all sensory studies were abnormal and all but one autonomic studies were abnormal. Generalized autonomic dysfunction was seen in four of the five patients tested who were clinically symptomatic. Four patients showed impairment in postganglionic sudomotor function, with abnormal adrenergic responses in one, and cardiovagal impairment in another (Table 4). When compared to DLRPN, there were no significant differences in NCS, quantitative sensory or autonomic testing. Needle examination showed more involvement of the L5, S1 myotomes in the painless group as compared to DLRPN which showed greater L2,3,4 involvement. Both cohorts had clear paraspinal denervation supporting proximal axonal pathology.
Table 3.
Electromyography in Painless Diabetic Motor Neuropathy and Painful DLRPN
| Painless Syndrome | Painful DLRPN | |||||||
|---|---|---|---|---|---|---|---|---|
| Nerve Conduction Studies** | ||||||||
| Nerve | Attribute | N | Median | Range | N | Median | Range | p* |
| Sural | SNAP, uV | 23 | 0.0 | 0.0 – 7.0 | 37 | 0.0 | 0.0 – 7.0 | NS |
| CV, m/s | 3 | 41.0 | 36.0 – 48.0 | 4 | 39.5 | 35.0 – 41.0 | NS | |
| DL, ms | 5 | 4.8 | 4.1 – 4.9 | 6 | 4.4 | 4.0 – 4.8 | NS | |
| Peroneal | CMAP, mV | 20 | 0.1 | 0.0 – 2.1 | 38 | 0.2 | 0.0 – 7.5 | NS |
| CV, m/s | 10 | 35.0 | 23.0 – 42.0 | 27 | 36.0 | 24.0 – 46.0 | NS | |
| DL, ms | 10 | 5.8 | 4.7 – 15.5 | 27 | 5.6 | 4.0 – 14.1 | NS | |
| Tibial | CMAP, mV | 22 | 0.7 | 0.0 – 5.5 | 31 | 0.7 | 0.0 – 11.4 | NS |
| CV, m/s | 14 | 34.5 | 20.0 – 43.0 | 25 | 36.0 | 20.0 – 48.0 | NS | |
| DL, ms | 14 | 5.7 | 4.6 – 9.4 | 25 | 5.2 | 3.8 – 10.8 | NS | |
| Ulnar | CMAP, mV | 23 | 4.8 | 0.7 – 9.5 | 33 | 6.8 | 0.0 – 16.1 | NS |
| CV, m/s | 23 | 47.0 | 24.0 – 61.0 | 29 | 48.0 | 35.0 – 57.0 | NS | |
| DL, ms | 23 | 3.3 | 2.5 – 9.9 | 29 | 3.2 | 2.2 – 5.2 | NS | |
| Median | SNAP,µV | 23 | 6.0 | 0.0 – 20.0 | 31 | 7.0 | 0.0 – 49.0 | NS |
| CV, m/s | 11 | 50.0 | 43.0 – 57.0 | 23 | 52.0 | 45.0 – 61.0 | NS | |
| DL, ms | 15 | 4.0 | 3.1 – 5.8 | 26 | 3.9 | 2.5 – 5.4 | NS | |
| Needle Studies† | ||||||||
| Attribute | N | Median | Range | N | Median | Range | p* | |
| L2,3,4 muscles | ||||||||
| Fibrillation potentials | 44 | 1.0 | 0 – 3 | 24 | 1.0 | 0 – 3 | NS | |
| Long MUPs | 39 | 1.0 | 0 – 3 | 22 | 2.0 | 1 – 3 | 0.0002 | |
| L5,S1 muscles | ||||||||
| Fibrillation potentials | 56 | 1.4 | 0 – 2.8 | 32 | 1.4 | 0 – 2.5 | NS | |
| Long MUPs | 55 | 1.1 | 0 – 2 | 28 | 1.5 | 0 – 3 | 0.04 | |
| Paraspinal muscles | ||||||||
| Fibrillation potentials | 35 | 1.5 | 0 – 3 | 22 | 1.0 | 0 – 2 | NS | |
| Long MUPs | 18 | 1.0 | 0 – 2 | 10 | 1.0 | 0 – 1 | NS | |
DLRPN = diabetic lumbosacral radiculoplexus neuropathy; NS = not significant (p > 0.05); SNAP = sensory nerve action potential; CV = conduction velocity; DL = distal latency; CMAP = compound muscle action potential; MUPs = motor unit potentials
Wilcoxon rank sum test for continuous data and Fisher's exact test for dichotomous data.
Patients were counted twice if both sides were sampled.
Values listed are median per level on one side, and are recorded as: 0=normal, 0.5=+/−, 1=+, 2=++, 3=+++ and 4=++++. Absent responses are not included.
Table 4.
Quantitative Sensory and Autonomic Testing in Painless Diabetic Motor Neuropathy and Painful DLRPN
| Painless Syndrome | Painful DLRPN | |||||||
|---|---|---|---|---|---|---|---|---|
| Quantitative Sensory Testing | ||||||||
| Modality | Site | N | Normal | Abnormal | N | Normal | Abnormal | p* |
| VDT | Foot | 9 | 1 | 8 | 17 | 4 | 13 | NS |
| Leg | 2 | 1 | 1 | 8 | 3 | 5 | NS | |
| Thigh | 1 | 0 | 1 | 2 | 1 | 1 | NS | |
| Any site | 9 | 1 | 8 | 17 | 4 | 13 | NS | |
| CDT | Foot | 8 | 3 | 5 | 17 | 8 | 9 | NS |
| Leg | 1 | 1 | 0 | 3 | 0 | 3 | NS | |
| Thigh | 1 | 1 | 0 | 2 | 1 | 1 | NS | |
| Any site | 8 | 3 | 5 | 17 | 8 | 9 | NS | |
| HP5 † | Foot | 8 | 2 | 6 | 17 | 9 | 8 | NS |
| Leg | 2 | 0 | 2 | 6 | 2 | 4 | NS | |
| Thigh | 1 | 0 | 1 | 3 | 2 | 1 | NS | |
| Any site | 8 | 2 | 6 | 17 | 9 | 8 | NS | |
| Autonomic reflex screen (CASS score) | ||||||||
| Component | N | Median | Range | N | Median | Range | p* | |
| Sudomotor Index | 9 | 2.0 | 0.0 – 3.0 | 14 | 2.5 | 1.0 – 3.0 | NS | |
| Adrenergic Index | 9 | 2.0 | 0.0 – 4.0 | 14 | 2.5 | 0.0 – 4.0 | NS | |
| Cardiovagal Index | 9 | 2.0 | 0.0 – 3.0 | 13 | 2.0 | 0.0 – 3.0 | NS | |
| Total CASS | 9 | 6.0 | 0.0 – 10.0 | 14 | 7.0 | 2.0 – 10.0 | NS | |
DLRPN = diabetic lumbosacral radiculoplexus neuropathy; NS = not significant (p > 0.05); VDT = vibration detection threshold; CDT = cooling detection threshold; HP5 = Heat-Pain 5 threshold
Wilcoxon rank sum test for continuous data and Fisher's exact test for dichotomous data.
For HP5, abnormaility is <= 5th percentile or >= 95th percentile.
Pathological findings
Nerve biopsies showed evidence of ischemic injury and microvasculitis (Table 4). Overall, the density of myelinated nerve fibers was reduced in all but one biopsy, and was severe in eight. Evidence for nerve ischemia included: 1) focal or multifocal fiber loss (figure 1A, 1B); 2) focal perineurial degeneration or thickening (figure 2A); 3) epineurial neovascularization (figure 2A); 4) abortive regeneration of nerve fibers beyond the original perineurium forming microfasciculi (injury neuroma) (figures 1C, 2B); 5) regenerating clusters in 16 nerves (few in 11 and frequent in five). Onion-bulbs were seen frequently in one biopsy (figure 1D), occasionally in another, and were absent in all other biopsies.
Figure 1.
Figure 2.
Perivascular and vascular inflammation of small arterioles, venules and capillaries without involvement of larger vessels was encountered in all nerves, mostly affecting epineurial vessels, but occasionally affecting the perineurial and endoneurial vessels. The inflammatory infiltrates typically consisted of small collections of mononuclear lymphocytes (Table 5). All inflammatory collections reacted to CD45 (leukocyte) and to a lesser extent CD68 (macrophage). Inflammatory cells were within the vessel wall in 15/23 nerve biopsies, suggesting microvasculitis. Destruction of vessel wall elements with separation of the smooth muscle layers of the tunica media by leukocytes (features diagnostic of microvasculitis) were found in three nerves (Fig 2, C–H). Evidence of previous bleeding was noted in 11/23 nerves, frequently in the perineurium, often adjacent to microvessels. Teased fibers demonstrated increased frequency of axonal degeneration in 16 nerves (mild or moderate in 15 and severe in one) and an increased number of empty nerve strands (Table 5). A low rate of segmental demyelination that tended to be clustered along the length of individual myelinated nerve fibers was also seen. The painless diabetic motor neuropathy nerves were indistinguishable from painful DLRPN nerves demonstrating ischemic injury and microvasculitis. As compared to the painless cohort, both CIDP and DSPN nerve biopsies showed significantly less multifocal fiber loss, neovascularization, perineurial thickening, injury neuroma, axonal degenation, hemosiderin, and vessel wall inflammation. Neither cohort was suggestive of microvasculitis. DSPN but not CIDP nerves showed significantly less inflammation overall and CIDP biopsies exhibited significantly more segmental demyelination and onion-bulbs compared to all groups (Table 5).
Table 5.
Pathologic Findings of Biopsy in Painless Diabetic Motor Neuropathy, Painful DLRPN, CIDP and Diabetic Polyneuropathy
| Painless Syndrome | Painful DLRPN | CIDP | DSPN | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| N = 23 | N = 33 | p* | N = 25 | p* | N = 21 | p* | |||||
| Paraffin and Epoxy Sections | |||||||||||
| Endoneurial and Perineurial Abnormality | |||||||||||
| Fiber degeneration or loss | 22 | 25 | NS | 15 | 0.005 | 15 | 0.04 | ||||
| Multifocal fiber degeneration or loss | 11 | 19 | NS | 2 | 0.003 | 2 | 0.01 | ||||
| Focal perineurial degeneration | 3 | 6 | NS | 0 | NS | 0 | NS | ||||
| Focal perineurial thickening | 18 | 24 | NS | 2 | <0.0001 | 2 | <0.0001 | ||||
| Injury neuroma ‡ | 11 | 12 | NS | 0 | <0.0001 | 0 | 0.0002 | ||||
| Onion Bulbs | 2 | 1 | NS | 14 | 0.02 | 2 | NS | ||||
| Interstitial Abnormality | |||||||||||
| Perivascular inflammation | 23 | 33 | NS | 25 | NS | 6 | <0.0001 | ||||
| Individual cells (< 10 cells) | 2 | 0 | 3 | 5 | |||||||
| Small collections (11–50 cells) | 16 | 21 | 20 | 1 | |||||||
| Moderate collections (51–100 cells) | 3 | 7 | 1 | 0 | |||||||
| Large collections (> 100 cells) | 2 | 5 | 1 | 0 | |||||||
| Inflammation of vessel wall | 15 | 15 | NS | 2 | <0.0001 | 0 | <0.0001 | ||||
| Diagnostic of microvasculitis | 3 | 2 | NS | 0 | NS | 0 | NS | ||||
| Hemosiderin in macrophages | 11 | 19 | NS | 3 | 0.009 | 0 | 0.0002 | ||||
| Neovascularization | 17 | 21 | NS | 3 | <0.0001 | 1 | <0.0001 | ||||
| Teased Fiber Conditions § | Median | Range | Median | Range | p | Median | Range | p | Median | Range | p |
| Normal: A, B (%) | 52.2 | 13 – 100 | 45.7 | 0 – 92 | NS | 60 | 4 – 91 | NS | 77.6 | 62 –92 | <0.0001 |
| Demyelination: C, D (%) | 3.9 | 0 – 33 | 2.6 | 0 – 25 | NS | 23.1 | 6 – 58 | <0.0001 | 1.6 | 0 – 12 | NS |
| Remyelination: F, G (%) | 15.7 | 0 – 48 | 11.9 | 0 – 33 | NS | 11.5 | 0 – 32 | NS | 12.0 | 2 – 22 | NS |
| Axonal: E, H (%) | 12.5 | 0 – 67 | 30.9 | 0 – 100 | NS | 3.1 | 0 – 59 | 0.001 | 3.3 | 0 – 23 | 0.001 |
| Classifiable (no.) | 54.0 | 6 – 174 | 63.0 | 7 – 127 | NS | 78 | 27 – 100 | NS | 75 | 21 – 203 | 0.04 |
| Empty (no.) | 43.0 | 7 – 99 | 42.5 | 5 – 93 | NS | 32 | 4 – 27 | NS | 27 | 1 – 70 | NS |
DLRPN = diabetic lumbosacral radiculoplexus neuropathy; CIDP = chronic inflammatory demyelinating polyradiculoneuropathy; DSPN = diabetic polyneuropathy; NS = not significant (p > 0.05).
Wilcoxon rank sum test for continuous data and Fisher's exact test for dichotomous data.
Healthy control age (median 61.5 years) and gender (9M 5F) not significantly different than painless LE syndrome.
Injury neuroma = abortive regenerative activity outside of the original perineurium.
Teased fiber types: A = normal, B = myelin wrinkling, C = demyelination, D = demyelination and remyelination, E = axonal degeneration, F = remyelination, G = myelin reduplication, H = regeneration after axonal degeneration, empty = unclassifiable strands without myelinated fibers.
Treatment and Course
16/23 patients returned for follow-up at our institution, 13 of whom were treated with various immunotherapies; intermittent intravenous immunoglobulin (6), intermittent plasma exchange (3), and pulse intravenous methylprednisolone (4). 11/13 patients had detailed examinations at follow-up with overall significant improvement in NIS scores (Figure 3). Mean pre-treatment NIS score was 65.5 (range 16.5 to 131) and mean post-treatment 40.8 (range 11 to 74.5) (p value 0.018). Overall, symptoms and findings improved with time and the disease course seemed to be monophasic.
Figure 3.
DISCUSSION
There has been debate as to what is the cause of painless, motor and lower limb predominant neuropathy in diabetic patients. One possibility is that these cases represent one end of the spectrum of DLRPN. On the other hand, the symmetrical and painless characteristics of this motor predominant neuropathy may suggest that it is “diabetic CIDP” 22–24, 33. We recently studied the epidemiology of CIDP asking whether CIDP occurs more commonly in DM and found no association26. However, we acknowledge that if “diabetic CIDP” exists, it may manifest as this painless diabetic motor predominant neuropathy.
We found that the clinical features of painless diabetic motor neuropathy were slightly different than those of typical DLRPN. The painless syndrome tends to begin insidiously (none had acute onset) and progresses at a slower pace. Compared to DLRPN, many more had bilateral onset, whereas DLRPN commonly begins asymmetrically and unilateral. Overall, the clinical course was more severe and widespread than DLRPN as evidenced by more severe distal segment impairment (worse NIS leg and EMG changes in L5/S1 segments) and more frequent and severe upper limb involvement. However, most clinical characteristics were quite similar between these two conditions. Both are motor and lower limb predominant neuropathic syndromes with autonomic and sensory involvement of large and small fibers. Both are frequently associated with substantial weight loss. Additionally, the NCS/EMG findings demonstrate axonal (as opposed to demyelinating) radiculoplexus neuropathies with widespread but patchy changes and prominent denervation involving lumbosacral greater than thoracic or cervical segments in both conditions.
There were some features of the painless motor predominant diabetic neuropathy that resemble typical CIDP, such as the generalized symmetrical proximal and distal involvement, the slow course and the absence of pain. Elevated CSF protein is typical of all three groups, so is not useful in distinguishing painless patients from CIDP or DLRPN. The pattern of sensory involvement seen in the painless cohort was pan-modal (both large and small fiber), whereas CIDP is large fiber predominant34, 35. The prominent and severe autonomic involvement in the painless syndrome would be atypical for CIDP36. Although our patients had mild slowing of conduction velocities, they did not meet AAN CIDP electrophysiological criteria27 nor our criteria for CIDP26. As a result, we do not think that this painless diabetic motor neuropathy clinically or physiologically is correctly categorized as “diabetic CIDP.”
From the outset, we felt that the best way to classify the painless lower-limb motor predominant neuropathy would be by the pathophysiological disease mechanism. Proximal fascicular biopsies from mixed motor and sensory nerves would have been ideal but were not done. Given the retrospective nature of this study and the fact that patients had sensory involvement, we felt that distal sensory nerve biopsies were appropriate and would be informative. We found very consistent findings on the nerve biopsies demonstrative of ischemic injury from microvasculitis that was indistinguishable from typical DLRPN. Although segmental demyelination was observed on teased nerve fiber preparations, the rates were low and tended to be grouped demyelination, supporting secondary segmental demyelination from axonal atrophy37. Grouped demyelination was also seen in painful DLRPN associated with ischemic injury18. Only one biopsy raised the possibility of inflammatory demyelination with frequent onion-bulbs, but in this case the electrophysiology was axonal. In contrast, biopsies from CIDP patients did not show evidence of ischemic injury, hemosiderin-laden macrophages or vessel wall inflammation (microvasculitis). The rates of segmental demyelination and onion-bulbs were much higher and axonal degeneration much lower in CIDP than in the painless cohort. This data is strong evidence that the pathology of CIDP and painless, diabetic motor neuropathy are distinct and that they are not the same disease.
One may ask whether a concurrent or prominent DSPN is causing the sensory symptoms and the pathological findings with the proximal weakness resulting from a separate process. While we acknowledge that a minor component of painless motor predominant neuropathy may be due to DSPN, we contend that DSPN is unlikely a major player. Previous work has shown a strong correlation between the development of diabetic neuropathy, retinopathy and nephropathy after many years of hyperglycemic exposure38. Only 4/23 of our patients had chronic hyperglycemic complications. The median duration of DM in our patients was 5.4 years compared to 12.0 years in a community diabetic cohort. Furthermore, the widespread sensory and autonomic findings in our painless patients and in DLRPN have also been reported in non-diabetic LRPN patients39 supporting that these abnormalities are not primarily due to the diabetic state. Most significantly, when we compared the biopsies of our painless patients to those with DSPN there was significantly more ischemic injury, inflammation and microvasculitis (Table 5). Therefore, we judge that the clinical, electrophysiological and pathological findings in these painless patients are not due to metabolic complications from DSPN.
Given the retrospective nature of our study, the treatment response data of painless diabetic motor neuropathy patients needs to be interpreted with caution. The observation that almost all of those whom received immunotherapy had some improvement, and that the degree of improvement was significant, (as judged by NIS) supports an autoimmune pathogenesis. Three non-treated patients spontaneously improved as well. Overall, this data supports an inflammatory neuropathy that is likely treatment responsive and probably monophasic, as after the treatment course, the patients did not worsen or relapse. This clinical picture is more in keeping with a self-limited process such as DLRPN as opposed to the usual progressive course of CIDP.
Consequently, we conclude that the painless, lower-limb, motor predominant neuropathy seen in diabetic patients is “painless DLRPN” and represents a variant of DLRPN and not diabetic CIDP or a separate entity. Painless DLRPN patients may be more akin to those described by Chokroverty with a symmetrical, slowly progressive disease course12. The lack of pain may in fact be a partial explanation for the more insidious, slower course as these patients are unable to precisely date the onset of symptoms. As proposed by Asbury, we demonstrate that DLRPN is a spectrum of disorders exhibiting a range of clinical presentations. However, we do not find that some forms of DLRPN are due to metabolic factors whereas other forms are due to ischemic injury. Here, we provide evidence that ischemic injury from altered immunity and microvasculitis is the pathophysiological mechanism accounting for both painful and painless DLRPN.
Acknowledgments
Supported in part by a grant from the National Institute of Neurological Disease and Stroke (NS 36797). Dr. Garces-Sanchez was a fellow of the Spanish Institute of Health, Carlos III, and the Spanish Society of Neurology Foundation (CM03/00105).
P.A. Low, D.P. Polston
Footnotes
Disclosure: Authors report no conflicts of interest.
References
- 1.Dyck PJB, Norell JE, Dyck PJ. Microvasculitis and ischemia in diabetic lumbosacral radiculoplexus neuropathy. Neurology. 1999;53:2113–2121. doi: 10.1212/wnl.53.9.2113. [DOI] [PubMed] [Google Scholar]
- 2.Garland H, Taverner D. Diabetic myelopathy. British medical journal. 1953;1:1405–1408. doi: 10.1136/bmj.1.4825.1405. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Garland H. Diabetic amyotrophy. British medical journal. 1955;2:1287–1290. doi: 10.1136/bmj.2.4951.1287. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Barohn RJ, Sahenk Z, Warmolts JR, Mendell JR. The Bruns-Garland syndrome (diabetic amyotrophy). Revisited 100 years later. Archives of neurology. 1991;48:1130–1135. doi: 10.1001/archneur.1991.00530230038018. [DOI] [PubMed] [Google Scholar]
- 5.Raff MC, Asbury AK. Ischemic mononeuropathy and mononeuropathy multiplex in diabetes mellitus. The New England journal of medicine. 1968;279:17–21. doi: 10.1056/NEJM196807042790104. [DOI] [PubMed] [Google Scholar]
- 6.Asbury AK. Proximal diabetic neuropathy. Annals of neurology. 1977;2:179–180. doi: 10.1002/ana.410020302. [DOI] [PubMed] [Google Scholar]
- 7.Bradley WG, Chad D, Verghese JP, et al. Painful lumbosacral plexopathy with elevated erythrocyte sedimentation rate: a treatable inflammatory syndrome. Annals of neurology. 1984;15:457–464. doi: 10.1002/ana.410150510. [DOI] [PubMed] [Google Scholar]
- 8.Bastron JA, Thomas JE. Diabetic polyradiculopathy: clinical and electromyographic findings in 105 patients. Mayo Clinic proceedings. 1981;56:725–732. [PubMed] [Google Scholar]
- 9.Said G, Lacroix C, Lozeron P, et al. Inflammatory vasculopathy in multifocal diabetic neuropathy. Brain. 2003;126:376–385. doi: 10.1093/brain/awg029. [DOI] [PubMed] [Google Scholar]
- 10.Chokroverty S, Reyes MG, Rubino FA. Bruns-Garland syndrome of diabetic amyotrophy. Transactions of the American Neurological Association. 1977;102:173–177. [PubMed] [Google Scholar]
- 11.Sander HW, Chokroverty S. Diabetic amyotrophy: current concepts. Seminars in neurology. 1996;16:173–178. doi: 10.1055/s-2008-1040973. [DOI] [PubMed] [Google Scholar]
- 12.Chokroverty S, Reyes MG, Rubino FA, Tonaki H. The syndrome of diabetic amyotrophy. Annals of neurology. 1977;2:181–194. doi: 10.1002/ana.410020303. [DOI] [PubMed] [Google Scholar]
- 13.Said G, Goulon-Goeau C, Lacroix C, Moulonguet A. Nerve biopsy findings in different patterns of proximal diabetic neuropathy. Annals of neurology. 1994;35:559–569. doi: 10.1002/ana.410350509. [DOI] [PubMed] [Google Scholar]
- 14.Thomas PK, Eliasson SG. Diabetic Neuropathy. In: Dyck PJ, Thomas PK, Lambert EH, Bunge R, editors. Peripheral Neuropathy. 2nd ed. Vol. 2. Philadelphia: W.B. Saunders Company; 1984. p. 1775. [Google Scholar]
- 15.Said G, Elgrably F, Lacroix C, et al. Painful proximal diabetic neuropathy: inflammatory nerve lesions and spontaneous favorable outcome. Annals of neurology. 1997;41:762–770. doi: 10.1002/ana.410410612. [DOI] [PubMed] [Google Scholar]
- 16.Kelkar P, Masood M, Parry GJ. Distinctive pathologic findings in proximal diabetic neuropathy (diabetic amyotrophy) Neurology. 2000;55:83–88. doi: 10.1212/wnl.55.1.83. [DOI] [PubMed] [Google Scholar]
- 17.Llewelyn JG, Thomas PK, King RH. Epineurial microvasculitis in proximal diabetic neuropathy. Journal of neurology. 1998;245:159–165. doi: 10.1007/s004150050197. [DOI] [PubMed] [Google Scholar]
- 18.Dyck PJB, Engelstad J, Norell J, Dyck PJ. Microvasculitis in non-diabetic lumbosacral radiculoplexus neuropathy (LSRPN): similarity to the diabetic variety (DLSRPN) Journal of neuropathology and experimental neurology. 2000;59:525–538. doi: 10.1093/jnen/59.6.525. [DOI] [PubMed] [Google Scholar]
- 19.Kawamura N, Dyck PJ, Schmeichel AM, et al. Inflammatory mediators in diabetic and non-diabetic lumbosacral radiculoplexus neuropathy. Acta neuropathologica. 2008;115:231–239. doi: 10.1007/s00401-007-0326-2. [DOI] [PubMed] [Google Scholar]
- 20.Amato AA, Barohn RJ. Diabetic Lumbosacral Polyradiculoneuropathies. Current treatment options in neurology. 2001;3:139–146. doi: 10.1007/s11940-001-0049-y. [DOI] [PubMed] [Google Scholar]
- 21.Sharma KR, Cross J, Ayyar DR, et al. Diabetic demyelinating polyneuropathy responsive to intravenous immunoglobulin therapy. Archives of neurology. 2002;59:751–757. doi: 10.1001/archneur.59.5.751. [DOI] [PubMed] [Google Scholar]
- 22.Gorson KC, Ropper AH, Adelman LS, Weinberg DH. Influence of diabetes mellitus on chronic inflammatory demyelinating polyneuropathy. Muscle & Nerve. 2000;23:37–43. doi: 10.1002/(sici)1097-4598(200001)23:1<37::aid-mus5>3.0.co;2-9. [DOI] [PubMed] [Google Scholar]
- 23.Krendel DA, Costigan DA, Hopkins LC. Successful treatment of neuropathies in patients with diabetes mellitus. Arch Neurol. 1995;52:1053–1061. doi: 10.1001/archneur.1995.00540350039015. [DOI] [PubMed] [Google Scholar]
- 24.Haq RU, Pendlebury WW, Fries TJ, Tandan R. Chronic inflammatory demyelinating polyradiculoneuropathy in diabetic patients. Muscle & Nerve. 2003;27:465–470. doi: 10.1002/mus.1250. [DOI] [PubMed] [Google Scholar]
- 25.Dyck PJ, Sherman WR, Hallcher LM, et al. Human diabetic endoneurial sorbitol, fructose, and myo-inositol related to sural nerve morphometry. Annals of neurology. 1980;8:590–596. doi: 10.1002/ana.410080608. [DOI] [PubMed] [Google Scholar]
- 26.Laughlin RS, Dyck PJ, Melton LJ, et al. Incidence and prevalence of CIDP and the association of diabetes mellitus. Neurology. 2009;73:39–45. doi: 10.1212/WNL.0b013e3181aaea47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Cornblath DR, Asbury AK, Albers JW, et al. Report from and ad hoc subcommittee of the American Academy of Neurology AIDS Task Force. Research Criteria for the diagnosis of chronic inflammatory demyelinating polyneuropathy (CIDP) Neurology. 1991;41:617–618. [PubMed] [Google Scholar]
- 28.Dyck PJ, Zimmerman I, Gillen DA, et al. Cool, warm, and heat-pain detection thresholds: testing methods and inferences about anatomic distribution of receptors. Neurology. 1993;43:1500–1508. doi: 10.1212/wnl.43.8.1500. [DOI] [PubMed] [Google Scholar]
- 29.Dyck PJ, Zimmerman IR, Johnson DM, et al. A standard test of heat-pain responses using CASE IV. Journal of the neurological sciences. 1996;136:54–63. doi: 10.1016/0022-510x(95)00277-9. [DOI] [PubMed] [Google Scholar]
- 30.Dyck PJ, Zimmerman IR, O'Brien PC, et al. Introduction of automated systems to evaluate touch-pressure, vibration, and thermal cutaneous sensation in man. Annals of neurology. 1978;4:502–510. doi: 10.1002/ana.410040605. [DOI] [PubMed] [Google Scholar]
- 31.Dyck PJ, Litchy WJ, Lehman KA, et al. Variables influencing neuropathic endpoints: the Rochester Diabetic Neuropathy Study of Healthy Subjects. Neurology. 1995;45:1115–1121. doi: 10.1212/wnl.45.6.1115. [DOI] [PubMed] [Google Scholar]
- 32.Low PA. Composite autonomic scoring scale for laboratory quantification of generalized autonomic failure. Mayo Clinic proceedings. 1993;68:748–752. doi: 10.1016/s0025-6196(12)60631-4. [DOI] [PubMed] [Google Scholar]
- 33.Sharma KR, Cross J, Farronay O, et al. Demyelinating neuropathy in diabetes mellitus. Archives of Neurology. 2002;59:758–765. doi: 10.1001/archneur.59.5.758. [DOI] [PubMed] [Google Scholar]
- 34.Hahn AF, Hartung H, Dyck PJ. Chronic Inflammatory Demyelinating Polyradiculopathy. In: Dyck PJ, Thomas PK, editors. Peripheral Neuropathy. Fourth ed. Vol. 2. Philadelphia: Elsevier Saunders; 2005. pp. 2221–2255. [Google Scholar]
- 35.Dyck PJ, Lais AC,MO, et al. Chronic Inflammatory Polyradiculoneuropathy. Mayo Clinic Proceedings. 1975;50:621–637. [PubMed] [Google Scholar]
- 36.Figueroa J, Dyck PJB, Low P. Autonomic Dysfunction in Chronic Inflammatory Demyelinating Polyradiculoneuropathy. Neurology. 2010;74 doi: 10.1212/WNL.0b013e3182494d66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Dyck PJ, Johnson WJ, Lambert EH, O'Brien PC. Segmental demyelination secondary to axonal degeneration in uremic neuropathy. Mayo Clinic proceedings. 1971;46:400–431. [PubMed] [Google Scholar]
- 38.Dyck PJ, Kratz KM, Karnes JL, et al. The prevalence by staged severity of various types of diabetic neuropathy, retinopathy, and nephropathy in a population-based cohort: the Rochester Diabetic Neuropathy Study. Neurology. 1993;43:817–824. doi: 10.1212/wnl.43.4.817. [DOI] [PubMed] [Google Scholar]
- 39.Dyck PJB, Norell JE, Dyck PJ. Non-diabetic lumbosacral radiculoplexus neuropathy: natural history, outcome and comparison with the diabetic variety. Brain. 2001;124:1197–1207. doi: 10.1093/brain/124.6.1197. [DOI] [PubMed] [Google Scholar]