DIABETICMedicine

DOI: 10.1111/j.1464-5491.2012.03598.x

Short Report: Complications

Asymmetrical attenuation of vibration sensation
in unilateral diabetic Charcot foot neuroarthropathy

J. Valabhji1, R. C. Marshall2, S. Lyons3, L. Bloomfield3, D. Hogg4, P. Rosenfeld5

and C. M. Gabriel2
1
Departments of Diabetes and Endocrinology, 2Neurology, 3Podiatry, 4Orthotics and 5Orthopaedic Surgery, St Mary’s Hospital, Imperial College Healthcare NHS
Trust, London, UK

Accepted 20 January 2012

Abstract
Aims To further characterize the distal sensory neuropathy in subjects with unilateral diabetic Charcot foot neuroarthr-
opathy.
Methods A retrospective cohort study to assess the level to which the sensory modalities of pinprick, light touch, vibration,
joint position and temperature were attenuated in the affected and unaffected limbs in subjects with unilateral Charcot. The
level to which the sensory modality was attenuated in each limb was assigned a score. The Wilcoxon signed rank test was
used to compare the scores in the affected and unaffected limbs and also to compare the scores of the different sensory
modalities in the affected and unaffected limbs.
Results Fifty subjects with unilateral Charcot foot neuroarthropathy were assessed. Mean age was 45  SD 6 years for the
17 subjects with Type 1 diabetes and 62  10 years for the 33 subjects with Type 2 diabetes. Duration of diabetes was
21  13 years, HbA1c was 70  19 mmol ⁄ mol [8.6  1.8 %] and 15 subjects (30%) required renal replacement therapy.
The level of attenuation of vibration sensation was more proximal in the affected compared with the unaffected limbs
(P = 0.002). Pinprick, light touch, joint position and temperature sensations were not different. Joint position sensation was
less attenuated bilaterally than the other sensory modalities.
Conclusions Asymmetrical attenuation of vibration sensation may predict the side that will develop a Charcot joint and
may suggest a more important role for vibration sense loss than loss of other sensory modalities in the pathophysiology of
Charcot.
Diabet. Med. 29, 1191–1194 (2012)
Keywords Charcot, diabetic foot, peripheral neuropathy, sensory loss

diabetes and unilateral Charcot foot neuroarthropathy by

Introduction
Charcot foot neuroarthropathy in diabetes is usually unilateral.
Previous studies characterizing the peripheral neuropathy
associated with Charcot neuroarthropathy in diabetes have
assessed peripheral sensory thresholds distally at specific points
on the foot and have not demonstrated differences in sensory
thresholds between the affected and unaffected limbs [1,2]. As
the severity of peripheral neuropathy progresses, the levels to
which sensory modalities can be demonstrated as abnormal
ascend more proximally up the legs. We aimed to further
characterize the distal sensory neuropathy in subjects with
Correspondence to: Dr Jonathan Valabhji, Department of Diabetes and
Endocrinology, St Mary’s Hospital, Imperial College Healthcare NHS Trust,
London W2 1NY, UK. E-mail: jonathan.valabhji@imperial.nhs.uk
assessing the level to which normal sensory perception was
attenuated in the affected and unaffected limbs.
Subjects and methods
We performed a retrospective cohort study to assess the level to
which the sensory modalities of pinprick, light touch, vibration,
joint position and temperature were attenuated in the affected
and unaffected limbs in subjects with a unilateral Charcot foot.
Subjects had been assessed between July 2004 and July 2009.
The multidisciplinary diabetes foot care team includes a neu-
rologist with a special interest in peripheral neuropathy and
only subjects who had been assessed by the same neurologist
(CMG) were included. Pinprick had been assessed with a
Neurotip (Owen Mumford Ltd, Oxford, UK), determining

ª 2012 The Authors.

Diabetic Medicine ª 2012 Diabetes UK 1191
DIABETICMedicine Unilateral Charcot neuroarthropathy • J. Valabhji et al.

whether sharpness was perceived or not; the presence or ab-
sence of light touch perception with cotton wool (with the
patient’s eyes shut); the presence or absence of vibration prer-
ception with a standard 128-Hz tuning fork at each joint;
proprioception using standard small movements at the joint;
and temperature using a cold (unused, in a standard air-con-
ditioned room) tuning fork. Most had acute Charcot neuro-
arthropathy on initial presentation, although some had
presented with ulceration and deformity associated with a
quiescent Charcot process. Subjects with bilateral Charcot
neuroarthropathy constituted 9% of the population seen in the
clinic with neuroarthropathy and were not included in this
study. Subjects in whom diabetes was not considered the pri-
mary cause of the peripheral neuropathy were excluded; other
aetiological factors had been alcohol, Charcot–Marie–Tooth
disease (hereditary sensory motor neuropathy type 1a), human
immunodeficiency virus, paraproteinaemia and rheumatoid
arthritis.
The most proximal level to which each sensory modality
was attenuated in each limb had been recorded prospectively
at the time of clinical assessment as: 0, normal; 1, attenuated at
the hallux; 2, at the toes; 3, to the forefoot level; 4, to the
midfoot level; 5, to the ankle; 6, to sock level; 7, to mid-shin;
8, to upper shin; 9, to the knee; 10, to above knee level; 11, to
mid-thigh level; 12, to the anterior superior iliac spine; 13, to
the umbilicus; or 14, to the costal margin. These constituted an
ordinal scale or ‘score’ of increasing severity from 0 to 14. To
summarize the data distribution characteristics, the level of
attenuation of each sensory modality was grouped into one of
three categories: normal or attenuation to forefoot level (scores
0–3); attenuation to midfoot, ankle or sock level (scores 4–6);
and attenuation to mid-shin or above (scores 7–14), and the
number (percentage) of subjects in each of the three categories
was then described. However, the original scores from 0 to 14
were used for the statistical analyses. The Wilcoxon signed
rank test was used to compare the scores in the affected and
unaffected limbs and also to compare the scores of the dif-
ferent sensory modalities in the affected and unaffected limbs.
The Mann–Whitney test was used to compare the scores in
those with more distal Charcot involvement (midfoot and
forefoot) to those with more proximal Charcot involvement
(hindfoot and ⁄ or ankle) and also to compare the scores of
those with Type 1 and Type 2 diabetes. Continuous variables
with normal and skewed distributions are expressed as means
(standard deviation) and medians (interquartile range),
respectively.
It has been established by the Caldicott Guardian that there
are no patient confidentiality or information governance issues
relating to the analysis of the outcome of routine clinical
management within the setting of the multidisciplinary diabetic
foot clinic and for the publication of anonymized data derived
from it.
Results
The study cohort comprised 50 subjects with unilateral
Charcot foot neuroarthropathy. Seventeen (34%) had Type 1
diabetes and 33 (66%) had Type 2 diabetes. Age was 45  SD
6 years for those with Type 1 diabetes and 62  10 years for
those with Type 2 diabetes. Duration of diabetes was
21  13 years, HbA1c was 70  19 mmol ⁄ mol [8.6  1.8 %]
and 15 subjects (30%) required renal replacement therapy
(eight on haemodialysis, one on continuous ambulatory peri-
toneal dialysis, two with renal transplants and four with
simultaneous pancreas kidney transplants). The Charcot pro-
cess involved midfoot in 36 subjects, midfoot and forefoot in
four, midfoot and hindfoot in one, midfoot and ankle in one,
hindfoot and ankle in one and ankle alone in seven subjects.
Therefore, 10 of the 50 subjects had more proximal involve-
ment of Charcot neuroarthropathy, including hindfoot or ankle
areas or both. The Charcot process involved the left foot in 26
and right foot in 24 subjects. At the time of neurological
assessment, 16 subjects were documented to have current
ulceration of the affected foot, mostly in association with qui-
escent Charcot neuroarthropathy that had previously caused
deformity.

Table 1 Distribution of attenuation of the sensory modalities of pinprick, light touch, vibration, joint position and temperature sensations in the affected
vs. unaffected limb in unilateral diabetic Charcot foot neuroarthropathy

Level to which sensory modality attenuated in Level to which sensory modality attenuated in
affected limb, n (%) unaffected limb, n (%)
Sensory Midfoot ⁄ ankle ⁄ Midshin Midfoot ⁄ ankle ⁄ Midshin
modality Normal ⁄ forefoot sock level and above Normal ⁄ forefoot sock level and above P-value*

Pinprick 11 (22) 20 (41) 18 (37) 14 (28) 17 (35) 18 (37) 0.417

Light touch 10 (20) 13 (26) 27 (54) 13 (26) 15 (30) 22 (44) 0.208
Vibration 9 (18) 24 (48) 17 (34) 14 (28) 24 (48) 12 (24) 0.002
Joint position 31 (62) 19 (38) 0 (0) 34 (69) 15 (31) 0 (0) 0.250
Temperature 10 (21) 10 (21) 27 (58) 12 (25) 11 (23) 25 (52) 0.375

*P-values are derived from comparisons of the original scores (0–14) in the affected and unaffected limbs. Of the 10 data sets, five were
complete, three had 49 out of 50 data points recorded, one had 48 out of 50 recorded and one had 47 out of 50 recorded.

ª 2012 The Authors.

1192 Diabetic Medicine ª 2012 Diabetes UK
 Original article
The level of attenuation of vibration sensation was more
proximal in the affected compared with the unaffected limbs
(P = 0.002); proportions in each category were 18, 48 and
34% vs. 28, 48 and 24%, respectively (Table 1). The level of
attenuation of vibration sensation in the affected limb was
not more proximal in those subjects with ankle or hindfoot
involvement compared with those with just midfoot or
forefoot involvement (P = 0.386); proportions in each cate-
gory were 30, 40 and 30% vs. 15, 50 and 35%. Pinprick,
light touch, joint position and temperature sensations were
not different in the affected compared with the unaffected
limbs (Table 1). Joint position sensation was attenuated to a
lesser extent bilaterally than the other sensory modalities in
both the affected and unaffected limbs (P < 0.0001 for all
comparisons); in no subject was joint position sensation
attenuated above the ankle in either limb, and the majority
had either no attenuation of joint position sensation, or
attenuation only at the hallux. There were no differences in
the levels to which any of the sensory modalities were
attenuated in those with Type 1 diabetes compared with
those with Type 2 diabetes in either the affected or unaf-
fected limbs.
Discussion
The prevalence of distal sensory neuropathy in patients with
diabetes is of the order of 30% [3], but why a minority develop
the profound changes of Charcot neuroarthropathy, and why
the process is usually unilateral, is unclear. We have further
characterized the distal sensory neuropathy in subjects with
diabetes and unilateral Charcot foot neuroarthropathy and
found that vibration sensation is attenuated to a greater extent
in the affected compared with the unaffected limb. In addition,
we have confirmed that joint position sensation is symmetri-
cally less affected than the sensory modalities of pinprick, light
touch, vibration and temperature sensation, a phenomenon
described previously in distal sensory neuropathy associated
with diabetes [4]. Distal sensory neuropathy in diabetes is
characteristically symmetrical; a recent study demonstrated
asymmetry to be uncommon neurophysiologically [5]. Our
study raises the possibility that asymmetrical attenuation of
vibration sensation may predict the side that will develop a
Charcot neuropathic joint and may suggest a more important
role for vibration sense loss than loss of other sensory modal-
ities in the pathophysiology of Charcot neuroarthropathy. It
has been demonstrated that vibration perception at the hallux
was abnormal in subjects with Charcot neuroarthropathy [2],
but more extensive loss on the affected side has not been
documented previously. A potential confounding variable is the
possibility that a Charcot foot transmits vibration less effi-
ciently than a normal foot—and thus the difference could
represent simple mechanical changes in bone structure or
inflammatory changes around the joint. However, the fact that
the level of attenuation of vibration sensation was not
even more proximal in those subjects with more proximal
ª 2012 The Authors.
Diabetic Medicine ª 2012 Diabetes UK
DIABETICMedicine
involvement of Charcot neuroarthropathy suggests that this is
not the case.
We used a clinical score that takes into account the distal to
proximal progression of attenuation of sensory modalities in
distal sensory neuropathy. While a number of scoring systems
have been validated for the clinical characterization of distal
sensory neuropathy, none have been validated for use in a
cohort of subjects with diabetic Charcot foot neuroarthropathy
and all rely on the presence or absence of sensory signs at
specific distal points on the foot [6,7].
Because of close working with a very active renal unit, the
proportion of subjects on renal replacement therapy is higher in
our cohort than proportions reported in other studies [8], so that
uraemia or concurrent immunosuppression therapy may have
influenced the characteristics of the distal sensory neuropathy.
We have previously reported different clinical features of Char-
cot neuroarthropathy in some of those on immunosuppression
therapy following renal transplantation, but not different clinical
features of the associated neuropathy [9]. The higher proportion
of those on renal replacement therapy may also have influenced
the relative proportions of those with Type 1 and Type 2 dia-
betes and the differences in age between them.
An inescapable weakness of the study was that the neurol-
ogist who conducted the clinical assessments was not blinded to
the affected side. However, the strong statistical difference in
the levels to which vibration was attenuated in the affected
compared with the unaffected limbs makes it less likely that
bias was influential.
It is not possible to assess degrees of severe sensory neurop-
athy neurophysiologically; once the sensory action potentials
have been lost, further measurement of these is of course not
useful to determine change in severity, and thermal thresholds
are regarded to be poorly sensitive to change. Our results might
suggest that the development of more sensitive clinical mea-
surements of sensory loss could help our understanding of
Charcot neuroarthropathy in diabetes.
Competing interests
Nothing to declare.
References
1 Guy RJ, Clark CA, Malcolm PN, Watkins PJ. Evaluation of thermal
and vibration sensation in diabetic neuropathy. Diabetologia 1985;
28: 131–137.
2 Stevens MJ, Edmonds ME, Foster AV, Watkins PJ. Selective neu-
ropathy and preserved vascular responses in the diabetic Charcot
foot. Diabetologia 1992; 35: 148–154.
3 Young MJ, Boulton AJ, MacLeod AF, Williams DR, Sonksen PH. A
multicentre study of the prevalence of diabetic peripheral neuropathy
in the United Kingdom hospital clinic population. Diabetologia
1993; 36: 150–154.
4 Said G, Baudoin D, Toyooka K. Sensory loss, pains, motor deficit
and axonal regeneration in length-dependent diabetic polyneuropa-
thy. J Neurol 2008; 255: 1693–1702.
1193
DIABETICMedicine
5 Lo YL, Leoh TH, Lim LL, Fook-Chong S, Ng YL, Ratnagopal P. The
laterality index in the evaluation of distal sensorimotor neuropathy.
J Clin Neuromuscul Dis 2008; 10: 18–21.
6 Meijer JW, van Sonderen E, Blaauwwiekel EE, Smit AJ, Groothoff
JW, Eisma WH et al. Diabetic neuropathy examination: a hierar-
chical scoring system to diagnose distal polyneuropathy in diabetes.
Diabetes Care 2000; 23: 750–753.
7 Bril V, Perkins BA. Validation of the Toronto Clinical Scoring Sys-
tem for diabetic polyneuropathy. Diabetes Care 2002; 25: 2048–
2052.
1194
Unilateral Charcot neuroarthropathy • J. Valabhji et al.
8 Sohn M-W, Lee TA, Stuck RM, Frykberg RG, Budiman-Mak E.
Mortality risk of charcot arthropathy compared with that of
diabetic foot ulcer and diabetes alone. Diabetes Care 2009; 32: 816–
821.
9 Valabhji J. Immunosuppression therapy posttransplantation can be
associated with a different clinical phenotype for diabetic charcot
foot neuroarthropathy. Diabetes Care 2011; 34: e135.
ª 2012 The Authors.
Diabetic Medicine ª 2012 Diabetes UK