doi:10.1093/brain/awh558 Brain (2005), 128, 2164–2174
Altered motor nerve excitability in end-stage
kidney disease
Arun V. Krishnan,
1,2
Richard K. S. Phoon,
3
Bruce A. Pussell,
3
John A. Charlesworth,
3
Hugh Bostock
4
and Matthew C. Kiernan
1,2
1
Institute of Neurological Sciences, Prince of Wales Hospital, Randwick, Sydney, Australia,
2
Prince of Wales Medical
Research Institute and Prince of Wales Clinical School, University of New South Wales,
3
Department of Nephrology,
Prince of Wales Hospital, Randwick, Sydney, Australia and
4
Sobell Department of Neurophysiology, Institute of Neurology,
Queen Square, London, UK
Correspondence to: Dr Matthew Kiernan, Prince of Wales Medical Research Institute, Barker Street, Randwick,
Sydney, NSW 2031, Australia
E-mail: M.kiernan@unsw.edu.au
Although multiple toxins have been implicated in the development of uraemic neuropathy, no causative agent
has been identified. In the present study, the excitability properties of lower limbmotor nerves in patients with
end-stage kidney disease treated with haemodialysis were measured before, during and after a standard 5 h
haemodialysis session, in an attempt to explore the pathophysiology of uraemic neuropathy. Compoundmuscle
action potentials were recorded from tibialis anterior and extensor digitorum brevis, following stimulation
of the common peroneal nerve in 14 patients. Measures of excitability were assessed in relation to changes
in serum levels of potential neurotoxins, including potassium, calcium, urea, uric acid, parathyroid hormone
and b-2-microglobulin. Before dialysis, measures of nerve excitability were significantly abnormal in the
patient group for axons innervating tibialis anterior and extensor digitorum brevis, consistent with axonal
depolarization: refractoriness was increased and superexcitability and depolarizing threshold electrotonus
were reduced. Pre-dialysis excitability abnormalities were strongly correlated with serum K
+
. Correlation
was also noted between the severity of symptoms and excitability abnormalities. Haemodialysis normalized
the majority of nerve excitability parameters. In conclusion, lower limb motor axons in uraemic patients
are depolarized before dialysis. The correlation between serum K
+
and excitability measures indicates that
hyperkalaemia is primarily responsible for uraemic depolarization, and a likely contributing factor to the
development of neuropathy.
Keywords: membrane potential; nerve excitability; potassium; threshold electrotonus; uraemic neuropathy
Abbreviations: b-2M = b-2-microglobulin; CMAP = compound muscle action potential; EDB = extensor digitorum brevis;
ESKD = end-stage kidney disease; NCS = nerve conduction study; NSS = neuropathy symptom score; PTH = parathyroid
hormone; SNAP = sensory nerve action potential; TA = tibialis anterior; TEd = depolarizing threshold electrotonus;
TEh = hyperpolarizing threshold electrotonus; T-NSS = total neuropathy symptom score
Received February 24, 2005. Revised April 10, 2005. Accepted May 12, 2005. Advance Access publication June 9, 2005
Introduction
Peripheral neuropathy in end-stage kidney disease (ESKD)
presents as a length-dependent, distal sensorimotor poly-
neuropathy with greater lower limb than upper limb
involvement (Bolton, 1980; Asbury, 1993). Previous studies
of neuropathy in ESKD have demonstrated prevalence rates
which vary from 60 to 100%, depending on the choice of
nerve segments, the indices measured and the number of
nerves studied (Nielsen, 1973; Bolton, 1980; Ackil et al.,
1981; Van den Neucker et al., 1998; Laaksonen et al., 2002).
The pathophysiology of uraemic neuropathy has not been
established. The finding that neurological complications of
renal failure may be improved by dialysis (Hegstrom et al.,
1962) and that patients receiving peritoneal dialysis had a
lower incidence of neuropathy than haemodialysis patients
gave rise to the ‘middle molecule hypothesis’ (Babb et al.,
1971). This hypothesis postulated that the higher rate of
neuropathy in patients on haemodialysis was secondary to
retention of toxic molecules in the middle molecular range
#
The Author (2005). Published by Oxford University Press on behalf of the Guarantors of Brain. All rights reserved. For Permissions, please email: journals.permissions@oupjournals.org

of 300–12 000 Da (Vanholder et al., 1994), given that these
substances were poorly cleared by haemodialysis membranes.
Examples of such molecules include parathyroid hormone
(PTH) and b-2-microglobulin (b-2M), the levels of which
are elevated in patients with ESKD (Vanholder et al., 1994).
The hypothesis, however, remains unproven and the toxicity
of a number of these middle molecules remains contentious
(Vanholder et al., 1994; Bostock et al., 2004).
Measurements of nerve excitability, which provide infor-
mation about membrane potential and biophysical properties
of peripheral axons (Bostock et al., 1998; Burke et al., 2001),
have been used to study peripheral nerves in patients with
neuropathy and have provided information about disease
pathophysiology (Cappelen-Smith et al., 2001; Kiernan et al.,
2001a, 2002a, 2005; Kanai et al., 2003; Nodera et al., 2004).
A preliminary study of motor nerve excitability in the upper
limb of patients with ESKD demonstrated membrane poten-
tial changes—specifically membrane depolarization before
haemodialysis (Kiernan et al., 2002b)—with subsequent
improvement in nerve excitability after dialysis. Given the
length-dependent predisposition of uraemic neuropathy, typ-
ically worse in the legs than in the arms, the present study
has focused on lower limb motor nerve excitability. The aim
of the study was to expand the original study by investigating
the excitability properties of lower limb motor axons, before,
during and after haemodialysis in patients with ESKD.
In addition, correlations were explored between excitability
changes and the clinical severity of neuropathy, related to
changes in the serum levels of potential uraemic toxins and
the severity of neuropathic symptoms.
Methods
Studies were undertaken on 14 patients with ESKD (8men, 6 women:
age range, 17–69 years; mean age, 50.3 years) receiving thrice-weekly
haemodialysis, using a biocompatible low-flux polysulfone mem-
brane (Fresenius, Bad Homburg, Germany). All patients were
dialysed against a K
+
concentration of 2 mmol/l. None of the
patients had a history of other illnesses known to cause neuropathy
such as diabetes or amyloidosis and there was no history of exposure
to neurotoxic medications, including immunosuppressive therapy.
The causes of ESKD in this group were glomerulonephritis
(9 patients), polycystic kidney disease (1), medullary cystic kidney
disease (2) and hypertensive vascular disease (2).
Patients gave informed consent to the procedures, which were
approved by the South East Sydney Area Health Service Human
Research Ethics Committee (Eastern Section) and the Committee
on Experimental Procedures Involving Human Subjects of the Uni-
versity of New South Wales. The studies were performed in accord-
ance with the Declaration of Helsinki.
A neurological history and examination were initially undertaken
and symptoms were graded using the neuropathy symptom score
(NSS) (Dyck et al., 1980, 1987, 1992; Laaksonen et al., 2002). Patients
were asked about the presence of motor symptoms in the limbs
(subset IB) and sensory symptoms, both negative (subset IIA) and
positive (subset IIB). Each symptom received a score of 1 and the
number of symptoms present in each subset was added to give a
total neuropathy symptom score (T-NSS). The maximum possible
T-NSS was 9.
Routine nerve conduction studies were undertaken in all patients.
Neurophysiological indices which had previously been shown to be
sensitive markers of uraemic neuropathy were evaluated (Ackil et al.,
1981; Laaksonen et al., 2002). Studies were undertaken on the sural,
tibial, common peroneal and superficial radial nerves using a Mede-
lec Synergy system (Oxford Instruments, Surrey, UK) and conven-
tional nerve conduction techniques (Burke et al., 1974; Kimura,
1983). Nerve stimulation was performed at a frequency of 1 Hz
for motor nerves and 2 Hz for sensory nerves. Motor amplitudes
were measured peak to peak and sensory amplitudes were measured
as an average of the rising and falling phase amplitudes. Latency was
measured to the onset of the compound potential. For sensory stud-
ies, a bipolar recording electrode configuration was used with a
standard interelectrode distance of 4 cm (Eduardo and Burke,
1988). For tibial nerve F-wave studies, the latency was recorded as
the mean of 10 responses following supramaximal stimulation of the
nerve at the medial malleolus.
The excitability properties of lower limb motor nerves in patients
with ESKD treated with haemodialysis were measured before, during
and 1 h after a standard 5 h haemodialysis session using a previously
described protocol (Kiernan et al., 2000; Krishnan et al., 2004).
Recordings were obtained from tibialis anterior (TA) and extensor
digitorum brevis (EDB), following stimulation of the peroneal nerve
at the fibular neck. Skin temperature was monitored close to the site
of stimulation for the duration of each study.
Serum electrolytes, urea, creatinine, PTH and b-2M
were measured at the time of the excitability studies. Kt/V, a standard
and commonly accepted measure of dialysis adequacy (Daugirdas,
1995, 2000), was also calculated according to the following formula,
where K is the dialyser clearance, t is the length of the dialysis session
(hours) and V is the urea distribution volume (litres), U1 is pre-
dialysis urea (mmol/l); U2 is post-dialysis urea (mmol/l, 1 h after
dialysis), BW is body weight, DBW is the change in body weight
following dialysis.
Kt=V ¼
ln U1=U2
0:008 · tðÞ
þ 4
3:5 · U1=U2 · DBW=BW: ð1Þ
The current required to produce the desired CMAP (compound
muscle action potential) amplitude was determined using a compu-
terized threshold-tracking program (QTRAC version 5.2a, Institute
of Neurology, Queen Square, London, UK, with multiple excitability
protocol TRONDXM2) that was run on a Pentium computer
(Kiernan et al., 2000). Recordings were amplified (gain 1000,
bandwidth 5–10 kHz) and digitized using an analogue-to-digital
(A/D) board (DT2812, Data Translation, Marlboro, MA, USA),
with a sampling rate of 10 kHz. Stimulus waveforms were converted
to current using a purpose-built isolated linear bipolar constant-
current stimulator.
Stimulus–response curves were generated for test stimuli of 0.2
and 1 ms duration (Fig. 1). The slope of the 1 ms stimulus–response
curve and the magnitude of the tracking ‘error’ (i.e. the difference
between measured response and target response) were used to
optimize the subsequent threshold tracking. The peak 1 ms response
was also used to calculate the target response (40% of the supra-
maximal CMAP response). The ratio between the stimulus–response
curves for the two different stimulus durations was used to calculate
rheobase (Burke et al., 2001) and the strength–duration time con-
stant (t
SD
) of motor axons of different thresholds using Weiss’s
formula (Weiss, 1901; Mogyoros et al., 1996).
Uraemic neuropathy Brain (2005), 128, 2164–2174 2165