The sympathetic skin response (SSR) was studied in 47 diabetic patients

selected for the presence of symptoms and clinical signs of peripheral neu-
ropathy and in 24 normal control subjects. The SSR was present in all
controls but was absent at the foot in 66% and at the hand in 27.7% of the
diabetic patients. Absence of the SSR failed to correlate with other electro-
physiologic parameters on routine nerve conduction and electromyographic
studies. Although absent SSR was more often found in patients with symp-
toms of autonomic dysfunction (P < 0.05),there was no correlation with any
specific symptoms of autonomic involvement. The SSR was frequently ab-
sent, at least in the foot, in those patients with abnormal cardiac beat-to-
beat variability (expiratoty:inspiratoy, E: I, ratio) and pupil cycle time (PCT).
In addition there was a good correlation between the amplitude of the SSR
and the value of the E:l ratio (r = 0.81, P < 0.001). The SSR may be a
valuable adjunct in the assessment of autonomic involvement in diabetic
neuropathy, but its sensitivity requires further evaluation.
MUSCLE & NERVE 10:711-716 1987

SYMPATHETIC SKIN RESPONSE IN

DIABETIC NEUROPATHY
BElTY SOLIVEN, MD, RICARDO MASELLI, MD, JONATHAN JASPAN, MD,
ANDREW GREEN, MD, HANK GRAZIANO, MICHELE PETERSEN,
and JEAN-PAUL SPIRE, MD

Autonomic neuropathy is known to result in sub-
stantial morbidity in diabetic patients. '." Conven-
tional electrophysiologic procedures do not ade-
quately assess the dcgree of autonomic nervous
system involvement. There recently has been re-
newed interest in a previously described technique,
known as the sympathetic skin response (SSR), in
which changes in voltage of the skin in response to
a variety of stimuli are T h e SSR was
originally thought to depend on the synchronized
activation of' sweat glancis," although a contribu-
tion of' an epidermal component to the response
has been proposed by Edelberg and Wright.' 111
the current investigation we evaluated the uselul-
ness of the SSK in diabetic patients with peripheral
neuropathy, most of whom had symptoms of auto-
From the Department of Neurology (Drs. Soliven, Maselli. Mr. Graziano.
Ms. Petersen, and Dr. Spire) and Medicine (Drs. Jaspan and Green),
Brain Research Institute. University of Chicago, Chicago, IL.
Acknowledgements: We are indebted lo Dr Barry Arnason and Dr. Robert
Lovelace for help and advice in the preparation of this paper, Dr. Avertano
Noronha for encouraging us. and Cindy Suilivan for the typing of this
manuscript.
Address reprint requests to Dr Soliven at the Department of Neurology,
University of Chicago, 5841 S Maryland Avenue, Chicago, IL 60637
Received for publication May 13, 1986; revised manuscript accepted for
publication October 15, 1986.
10148-639N1008/0711$04.00/6
1987 John Wiley 8 Sons. Inc
nomic involvement. We also attempted to correlate
the results of the SSK with findings obtained by
other more routine electrophysiological tests as well
as with the results of traditional tests of autonomic
function such as the expiratory: inspiratory (E: I)
ratio on electrocardiographyg and pupil cycle
time.7.12.1"
PATIENTS AND METHODS
Thcrc were 47 diabetic patients (30 male, 17 fe-
male, age 50 2 I7 years) with symptoms and signs
of' peripheral neuropathy, as determined by a his-
tory and physical examination, chosen for study.
Patients were otherwise unselected, except for a
willingness to be included in the study, which was
approved by the Clinical Investigation Committee
of the institution. Written informed consent was
obtained from all participants. Thcre were 37 of
the 47 diabetic patients who had symptoms ofauto-
nomic involvement, iricluding one or more of the
following: orzhostatic hypotension ( 13/47), gastro-
paresis (20/47), bladder dysfunction (4/47), and male
impotence (24/30).
The SSK was obtained with the subject lying in
a supine position. Standard EMC; disc electrodes
wcre covered with conducting paste and attached
to the palm (GI) and dorsum (G2) of the right hand
as well as to the sole (GI) and dorsum (G2) of the
right foot. Recordings were obtained on a T E C A

Sympathetic Skin Response MUSCLE & NERVE October 1987 711

 zw IJV
500 IM

HAND FOOT
FIGURE 1. SSR from a 27-year-old normal control subject.

TE4 electromyograph (TECA Gorp., Pleasantville,
NY) with a frequency response from 1.6 H z to 3.2
KHz. T h e amplification sensitivity was between 50
and 500 pvldivision, and the sweep speed was 500
msec/division. The stimulation consisted of square
pulses of 200 psec duration delivered to the skin
of the hand and foot. The skin temperature in all
patients was kept above 32" C. The SSR consists of
a biphasic potential with an initial negative peak,
varying in amplitude (peak to peak) from test to
test and is subject to habituation (Fig. 1).
In 25 of the diabetic patients, routine nerve
conduction velocity studies were performed. In 36
patients the expiratory/inspiratory ratio was deter-
mined on electrocardiography by dividing the
longest R-R interval following expiration by the
shortest R-R intervals during inspiration obtained
during steady breathing at 6 cycles/min.'A mean
ratio was calculated from eight consecutive expi-
ration/inspiration cycles. In 33 patients the pupil
cycle time was obtained using a slit-lamp as de-
scribed previously by others.'*'*.''Briefly, a 1/2 mm
thick beam of light is directed perpendicularly to
the inferior limbus of the iris. T h e beam is elevated
until it intersects the pupillary margin, eliciting a
constriction that shields the beam from the retina.
This causes the pupil to dilate, setting u p a cycle
of constriction dilation. There are 90 cycles timed
over 3 or 4 periods of 20-30 seconds each and the
mean pupil cycle time thereby obtained.
STATISTICAL METHODS
Statistical methods included chi-square analysis,
Student's t-tests, and coefficient of correlation where
appropriate.
RESULTS
The SSR was present in 100% of controls but was
absent in the foot in 31 of 47 (66%) of the diabetic
patients (P < 0.01), (Table 1, Fig. 1). Of the patients
with absent foot SSR, 28 (90%) had one or more
symptoms of autonomic involvement, whereas 3
(10%) had no autonomic symptoms. Conversely, of

Table 1. SSR in controls and diabetics.

Control Diabetics P value
Number of patients 24 47
Age 27-85 years (X = 42.3) 24-80 years (X = 50)
Duration since time 1 month to 58 years
of diagnosis
SSR absent in foot 0% 31/47 = 66% <0.01
SSR absent in 0 Yo 13/47 = 27.7%
hand
SSR present in 100% 16/47 = 34%
hand and foot
-
x latency
Hand 1.31 r 0.18 seconds 1.44 ? 0.09 seconds NS
Foot 1.93 f 0.17 seconds 1.99 f 0.18 seconds NS
jT amplitude
Hand 791.7 ? 346.6 pV 471.9 2 368 pV <0.005
Foot 388.3 2 227 pV 196.5 ? 138.3 pV <0.005
NS = not s!gn!f!cant

712 Sympathetic Skin Response MUSCLE & NERVE October 1987


351
SSR ABSENT SSR
AT LEAST IN PRESENT
THE FOOT
RGURE 2. SSR and autonomic symptoms.
the patients with autonomic symptoms, 28 (75.6%)
had absent SSR, whereas 9 (24.4%) had normal SSR
(Fig. 2). These differences were significant (P <
0.05). Thus, there appears to be an association be-
tween absence of the SSR, at least in the foot, and
the presence of autonomic neuro athy in diabetic
and any single autonomic symptom such as or-
E
patients. However, no correlation etween the SSK
thostatic hypotension, gastroparesis, bladder dys-
function, o r impotence was noted. Similarly, no as-
sociation was found between the SSR and patient
age, sex, duration, or type of diabetes (Table 2).
N o age and sex differences in the SSR were ob-
served in the control subjects.
The mean latencies of the foot and hand SSR
in the diabetic subjects were not significantly dif-
ferent from the controls. On the other hand, the
mean amplitude of the hand and foot SSR was
significantlyreduced compared with controls in those
diabetic patients in whom a response was detected.
Table 2. SSR and age of patients, duration of diabetes, and type of
~~
X age (years)
X duration of DM (years)
Type l/Type II (%)
Type I = insulin dependent.
Type I / = noninsulin dependent.
N S = not significant.
Sympathetic Skin Response
However there was overlap in amplitude between
normals and diabetics (Fig. 3). Among the 25 pa-
tients in whom nerve conduction velocities were
performed, absence of the SSR did not correlate
with nerve conduction velocities or amplitude of
nerve and muscle action potentials (Table 3).
The SSR was frequently abnormal in patients
with other objective signs of autonomic dysfunction
such as abnormal E:I ratio (79%) and abnormal
pupil cycle time (75.8%)(Tables 4 and 5).T h e value
of the E: I ratio correlated well with the amplitude
of foot SSR (r = 0.81, P < 0.001) and with that of
the hand SSR (r = 0.42, P < 0.05) (Fig. 4). On the
other hand there was no correlation between the
amplitude of the SSR in either hand or foot and
pupil cycle time.
DISCUSSION
The assessment of autonomic function poses a chal-
lenging problem because of the inherent limita-
tions of routine electrophysiologic methods that re-
flect the status of fast myelinated fibers only. Using
microneurography, Fagius and Wallin"' were able
to demonstrate dissociation of the vasoconstrictor
and sudomotor activity of the sympathetic path-
ways. One indirect method of studying the integrity
of the sudomotor fiber population is the psycho-
galvanic response,".'.':' which measures changes in
skin resistance in response to emotional influences.
T h e method used in this study measures changes
in voltage from the skin surface as a result of ac-
tivity in the sudomotor fiber triggered by electrical
stimulation. Its main advantage is that it can be
readily performed with routine electromyographic
equipment.
Our findings indicate that the SSR was easily
recorded and universally present in normal sub-
jects. It was found that this reflex was frequently
absent in patients with clinically evident diabetic
neuropathy, in agreement with previous re-
port~.'.'~These findings are consistent with the ob-
servations that the presence of the SSR depends
diabetes.
Normal SSR Abnormal SSR t-test
50.3 t 18.4 50.1 f 14.9 NS
20.4 t 16.6 16.5 f 10.9 NS
38.5/61.5% 355164.5% NS
MUSCLE & NERVE October 1987 713
 "I . HAND SSR FOOT SSR

mo

CONTROLS DIAEETICS

FIGURE 3. Amplitude of SSR in controls and diabetic patients. In the first five diabetic patients, SSR was simply measured as present
or absent, and for this reason, these data from these five patients is not presented in this figure.

Table 3. SSR and electrodiagnostic parameters in diabetics.

Normal hand SSR' Absent hand SSR' t-test
K ulnar SNAP (pV)t 3.87 f 3.89 (16) 1.67 2 2.69 (9) NS
(nl 2 5 pV)
K ulnar CMAP (mV)* 11.3 2 3.64 (16) 9.28 f 4.12 (9) NS
(nl 2 6 mV)
K ulnar SCV (m/sec) 48.3 2 ll.S(ll) 52.5 -t 3.18 (2)s NS
(nl = 44-71 m/sec)
X ulnar MCV (m/sec) 47.1 2 4.25 (16) 48.8 ? 5.27 (9) NS
(nl = 47-60 mlsec)
Normal foot SSR Absent foot SSR t-test
K peroneal CMAP (mV)* 3.03 f 1.9 (6) 2.34 2 2.57 (19) NS
(nl s 2.0 mV)
i?peroneal MCV (m/sec) 35.9 f 4.1 (6) 36.5 5 4.04 (14)s NS
(nl = 38-61 m/sec)
H reflex present 50% (3/6) 47.4% (9119) NS
NS = not significant
'Number of patrents in parenthesis.
tSensory nerve action potential.
$Compound muscle action potential measured from peak to peak.
Discrepancy in the number of subjects when calculating conduction velocity is due to absent ulnar
SNAP or peroneal CMAf in some patients. Such patients are excluded from calculatron.

on intact conduction through the postganglionic
unmyelinated fibers that are frequently affected in
diabetics. Although there was a tendency for the
amplitude of the SSR to be lower in diabetics than
in controls, as also noted by others,' the great var-
iability of the response and its overlap with controls
makes applitude unreliable as an index of abnor-
mality. For these reasons w e have considered the
SSK abnormal only when it is absent. Amplitude
variability is related to the responsiveness of the
sympathetic nervous system to various stimuli. T h e
absence of significant ditference in the latency of
the SSR between the normal subjects and diabetics
may be explained by the fact that large myelinated
fibers do not contribute to the SSR. Thus, if the
number of functional fibers is above a certain
threshold, the response will be elicited, albeit with
decreased amplitude, and if not, it will be absent.
Interestingly, although perhaps not surpris-
ingly, absence of the SSR did not correlate with the

Table 4. SSR and expiratory-inspiratory ratio in diabetics Table 5. SSR and pupil cycle time in diabetics
NL SSR ABN SSR Total NL SSR ABN SSR Total
NL E:l ratio (>1.15) 3 0 3 NL PCT (<920 msec) 1 3 4
ABN E:l ratio ( ~ 1 . 1 5 ) 7 26 33 ABN PCT (>920 msec) 7 22 29
Total 10 26 36 Total 8 25 33

714 Sympathetic Skin Response MUSCLE & NERVE October 1987

 HAND SSR
1600
1200
11V
800.
400,
.......-".. I 42
E /I Ratio
FIGURE 4. Correlation between amplitude of SSR and expiratory-inspiratory ratio
severity of the peripheral demyelination as mea-
sured by slowing of conduction velocity nor with
the severity of axonal damage as indicated by a
decreased amplitude of the compound muscle and
nerve action potentials. This is at variance with the
report by Shahani et al.15who found the SSR to be
more frequently abnormal in axonal neuropathies
than in demyelinating neuropathies. This may,
however, simply reflect differences in study pop-
ulations, since all of our patients were diabetics.
T h e SSR was abnormal more often in patients with
than without symptoms of' autonomic involvement
(P < 0.05), although no correlation with specific
symptoms was observed. However, it should be noted
that most of the autonomic symptoms such as gas-
troparesis, bladder atony, and to some extent im-
potence are due to parasympathetic dysfunction.
Although orthostatic hypotension is related to sym-
pathetic dysfunction, it involves vasoconstrictor re-
flexes mediated through fibers different from those
REFERENCES
I . Clarke BF, Ewing DJ: <:ardiovascular reflex tests in the
natural history of diabetic autonomic neuropathy. NY S h t e
J Med 82:903-908, 1982.
2. Edelberg R, Wright DJ: Two galvanic skin response effector
organs and their stimulus specificity. Pqch@hysiokiigf 1:39-47,
1964.
Ewing DJ, Campbell IW, Clarke HF: Assessment of cdrdio-
vascular effects in diabetic neuropathy and prognostic im-
plications. Ann Intern Med 92:308-31 I , 1980.
Fagius J. Wallin BG: Sympathetic reflex latencies and con-
duction velocities in normal man. ). Neurol Sci 47:433-448,
1980.
, .
Faeius 1. Wallin BG: Sympathetic reflex latencies and con-
dukonvelocities in patients with polyncuropathy.] Neurol
SCZ 473449-461, 1980.
Goadby HK, Downman CBB:Peripheral vascular and sweat
gland rcHexcs in diabetic neuropathy. Clin Sci Mol Med 45:
28 1-289, 3973.
Sympathetic Skin Response
FOOT SSR
500
400 ..
'" 300.
..
200.
100 I 81
E l l Ralio
associated with sudomotor a~tivity.~.~.'" T h e am-
plitude of the SSR correlated well with the expir-
atory: inspiratory ratio but not with pupil cycle time.
The lack of correlation with the latter may in part
reflect the fact that visual pathways participate in
the reflex arc of pupillary response, in addition to
the sympathetic and parasympathetic pathways. It
would be of interest to correlate the SSR with other
tests of sudomotor activity such as quantitation of
sweat gland output in response to acetylcholine ion-
tophoresis,".' ' since this represents another mea-
sure of peripheral autonomic function. We are cur-
rently pursuing this question.
Addendum: In a subsequent study of 18 patients
with more severe diabetic neuropathy, absence of
SSR appeared to correlate with signs of axonal in-
volvement such as reduced amplitude of sensory
nerve action potentials and/or compound muscle
action potentials.
7. Hamilton W. Drewry RD: Edge-light pupil cycle time and
optic nerve disease. Ann Ophfhalmol 15:7 14-72 I , 1983.
8. Knezevic W, Bajada S: Peripheral autonomic surface po-
tential: a quantitative technique for recording sympathetic
conduction in man. j Neurol Scz 67:239-251, 1985.
9. Lader MH, Montagu I D : T h e psychogalvanic reflex: a
pharmacological study of' the peripheral mechanism.] Neu-
rol Neurosurg Psychiatty 25: 126- 133, 1962.
10. Low PA, Caskey PE, Tuck KR: Quantitative sudomotor
axnn reflex test in normal and neuropathic subjects. Ann
Neurol 14:573-580, 1983.
I I . Macmillan AL, Spalding J M K : Human sweating response
to elcctrophoresed acetylcholine: a test of postganglionic
function. J Neurol Neurosurg Psvchzahy 32: 155-160, 1969.
12. Miller SD. Thompson HS: Edge-light pupil cycle time. Br
J Oph~halmol62:495-500, 1978.
IS. Prout BJ: Independence of the galvanic skin reflex f'roni
MUSCLE & NERVE October 1987 715
 the vasoconstrictor reflex in man. J Nrurol Nrurosurg P y -
clizufty 30:319-324, 1967.
14. Schilf E, Schubert A: Uber das sogenannte psychoglavan-
ische reflexphanomen beini frosch und seine beziehung
zurn vegetativen nerven system. Pflwgers Arch Physiol 195:
75, 1922.
716 Sympathetic Skin Response
15. Shahani BT, Halperin J]. Boulu P, Cohen 1: Synipathetic
skin response: a method of assessing unmyelinated axon
dysfunction in peripheral neuropathies.J Nrurol Nruro.surg
PsychiufT 47556-542, 1984.
16. Weinstein J M , Van Gilder JC, Thompson HS: Pupil cycle
time in optic nerve compression, Am,! Optlalmol89:263-267.
1980.
MUSCLE & NERVE October 1987