Acute Tetrodotoxin-Induced Neurotoxicity
after Ingestion of Puffer Fish
Matthew C. Kiernan, PhD, FRACP,
1,2
Geoffrey K. Isbister, MBBS, FACEM,
3
Cindy S.-Y. Lin, PhD,
1,2,4,5
David Burke, DSc, FRACP,
4
and Hugh Bostock, PhD, FRS
5
This study documents the effects of puffer-fish poisoning on peripheral nerve. Excitability measurements investigated
membrane properties of sensory and motor axons in four patients. The median nerve was stimulated at the wrist, with
compound muscle potentials recorded from abductor pollicis brevis and compound sensory potentials from digit 2.
Stimulus–responses, strength–duration time constant (

SD
), threshold electrotonus, and current–threshold relations were
recorded. The urine of each patient tested positive for tetrodotoxin. Compared with controls, axons were of higher
threshold, compound muscle action potentials and compound sensory nerve action potentials were reduced in amplitude,
latency was prolonged, and

SD
was reduced. In recovery cycles, refractoriness, superexcitability, and late subexcitability
were decreased. Threshold electrotonus of motor axons exhibited distinctive abnormalities with less threshold decline
than normal on depolarization and greater threshold increase on hyperpolarization (p < 0.0005 for each patient). The
changes in excitability were reproduced in a mathematical model by reducing sodium (Na

) permeabilities by a factor of
two. This study confirms that the neurotoxic effects of puffer-fish poisoning can be explained by tetrodotoxin blockade
of Na

channels. It demonstrates the ability of noninvasive nerve excitability studies to detect Na

channel blockade in
vivo and also the utility of mathematical modeling to aid interpretation of altered excitability properties in disease.
Ann Neurol 2005;57:339–348
Voltage-dependent sodium (Na


) channels underlie
action potential generation and are the chief determi-
nants of membrane excitability in human nerves.
1,2
Mammalian neurons generate heterogeneous Na


cur-
rents that can be pharmacologically discriminated by
their sensitivity to tetrodotoxin (TTX).
2,3
This marine
toxin blocks subtypes of Na


channels over the single
nanomolar concentration range. Na


currents are de-
fined as either TTX-sensitive (TTX-s) or TTX-resistant
(TTX-r), and they may be differentially expressed in
small and large neurons.
4,5
TTX-s low-threshold Na


currents contribute particularly to action potential gen-
eration and to the control of peripheral nerve excitabil-
ity.
6
TTX-s late current also plays a role in generating
pathological activity in the peripheral nervous system
by driving the membrane potential oscillations that
cause spontaneous activity in demyelinated axons.
7,8
In
contrast, TTX-r currents are preferentially expressed in
small neurons, supporting small-diameter axons,
9
many
of which are involved in nociception.
Whereas the in vitro effects of TTX are well charac-
terized, less is known about its in vivo effects in hu-
mans. TTX may be ingested accidentally or intention-
ally by consuming puffer fish. TTX is present in high
concentrations in the liver, ovaries, intestines, and skin
of these fish.
10
Fugu (puffer-fish fillets) is a delicacy,
and although processing of fugu is licensed, puffer-fish
poisoning accounts for more deaths than any other
type of food poisoning in Japan.
11
Despite this statis-
tic, there is limited information on nerve function in
patients after puffer-fish poisoning, and it is possible
that the clinical effects involve toxins other than or in
addition to TTX. Oda and colleagues
12
undertook se-
rial nerve conduction studies in a single patient and
documented reduced amplitudes of compound motor
and sensory potentials, slowing of conduction veloci-
ties, and prolongation of distal motor latencies and
F-wave latencies. Although Na


-channel blockade
could produce these changes, amplitude and conduc-
tion velocity are nonspecific indicators of pathophysi-
ology.
In contrast to nerve conduction studies, measure-
ments of axonal excitability are capable of providing an
indirect measure of resting membrane potential and
From the
1
Prince of Wales Medical Research Institute, University of
New South Wales;
2
Department of Neurology, Prince of Wales
Hospital;
3
Emergency Department, Newcastle Mater Misericordiae
Hospital and University of Newcastle;
4
Institute of Clinical Neuro-
sciences, University of Sydney and Royal Prince Alfred Hospital,
Sydney, Australia; and
5
Sobell Department of Neurophysiology, In-
stitute of Neurology, London, United Kingdom.
Received Jun 7, 2004, and in revised form Nov 16. Accepted for
publication Dec 7, 2004.
Published online Feb 24, 2005, in Wiley InterScience
(www.interscience.wiley.com). DOI: 10.1002/ana.20395
Address correspondence Dr Kiernan, Prince of Wales Medical Re-
search Institute, Barker Street, Randwick, Sydney, NSW 2031, Aus-
tralia. E-mail: m.kiernan@unsw.edu.au
© 2005 American Neurological Association 339
Published by Wiley-Liss, Inc., through Wiley Subscription Services

the activity of a variety of ion channels, energy-
dependent pumps, and ion exchange processes acti-
vated during impulse conduction.
2
This study was un-
dertaken to document the effects of acute neurotoxicity
after ingestion of puffer fish, with particular focus on
the changes in nerve excitability, to clarify whether all
physiological changes can be explained by TTX. In
these patients, significant changes were evident in Na


channel–dependent nerve excitability parameters.
Mechanistic insight was provided by reproduction of
the patient data using a mathematical model of the hu-
man axon in which transient and persistent Na


con-
ductances were reduced by a factor of two. We con-
cluded that TTX can explain all abnormalities and that
additional toxins need not be invoked.
Patients and Methods
Nerve excitability recordings were taken for four adult pa-
tients (three men and one woman; age range, 33–47 years)
from a group of nine patients who had eaten puffer fish.
Seven adults and 2 children consumed a soup made from
approximately 30 puffer fish, which had been gutted with
heads intact and boiled in fresh water. The clinical findings
for the group have been reported previously.
10
Each of the
patients had been previously well and took no regular med-
ications. Recordings were undertaken 21 hours after soup
consumption.
Case Histories
PATIENT 1. A 47-year-old woman experienced numbness
of her lips 1 hour after consuming the soup. The numbness
spread to her tongue, throat, and then hands and feet, as-
cending proximally in the upper limbs to the level of the
elbow. She became ataxic and described a drifting feeling and
a sensation of motion sickness. These symptoms progressed
over a 2-hour period, during which she felt nauseated and
vomited on multiple occasions. On examination, gait was
unsteady, and power was mildly reduced in a generalized
pattern. Reflexes were symmetrically normal, and there was
no sensory deficit at the time of neurophysiological testing.
PATIENT 2. A 47-year-old man experienced numbness of
the lips and tongue 30 minutes after consuming two bowls
of the soup. He described difficulty swallowing saliva and a
light-headed sensation. Numbness then developed in his fin-
gers and toes, and he was unable to hold objects. Examina-
tion indicated an ataxic gait, power was generally reduced,
and reflexes were preserved.
PATIENT 3. A 41-year-old man described the onset of
numbness involving his lips, teeth, and tongue 90 minutes
after consuming three fishtails and one bowl of soup. Thirty
minutes later, his fingertips became numb; within another
30 minutes, his toes were similarly affected. His gait became
unsteady, and he fell on a number of occasions. On exami-
nation, gait was ataxic, reflexes were absent, and sensation
was decreased in a “glove-and-stocking” distribution.
PATIENT 4. A 33-year-old man experienced numbness of
his lips approximately 1 hour after consuming a bowl of the
soup. His fingertips and toes later became numb, and he felt
as though he “could fly.” He described generalized weakness
and a restless sensation. On examination, he was ataxic and
areflexic.
These neurotoxic effects gradually resolved over the next
week, and all patients made a complete recovery. TTX was
detected in the urine of all four patients (see below).
Nerve Excitability Recordings
Studies were performed using a protocol designed to rapidly
measure a number of different nerve excitability parame-
ters.
13
Compound muscle action potentials (CMAPs) were
recorded from thenar muscles using surface electrodes over
abductor pollicis brevis, with the active electrode at the mo-
tor point and the reference on the proximal phalanx. Anti-
dromic compound sensory nerve action potentials were re-
corded from digit 2. The signals were amplified and digitized
by computer (486 PC) with A/D board (DT2812; Data
Translation, Marlboro, MA), using a sampling rate of
10kHz. The stimulus currents were with the active electrode
over the median nerve at the wrist, and the reference elec-
trode was about 10cm proximal over muscle. Stimulation
and recording were controlled by QTRAC software (version
5.2; Institute of Neurology, London, UK, with multiple ex-
citability protocol, TRONDXM).
Test current pulses of 0.2 or 1 millisecond for motor ax-
ons (or 0.1 or 0.5 millisecond for sensory axons) were ap-
plied at 0.8-second intervals and combined with suprathresh-
old conditioning stimuli or subthreshold polarizing currents,
as required. The amplitude of the CMAP was measured
from baseline to negative peak. Compound sensory nerve ac-
tion potential amplitudes were measured from negative peak
to the subsequent positive peak after baseline subtraction.
Skin temperature was monitored close to the stimulation site
and kept greater than 32°C.
The sequence of recordings has been described previous-
ly.
13,14
Stimulus–response curves were recorded separately for
test stimuli of two durations: 0.1 and 0.5 millisecond for
sensory axons (Fig 1A), and 0.2 and 1 millisecond for motor
axons (Fig 2A). The ratio between the stimuli of different
durations required to evoke the same response was used to
estimate rheobase and the strength–duration time constant
(
SD
) of axons of different thresholds (see Figs 1D and 2D).
Prolonged subthreshold currents were used to alter the po-
tential difference across the internodal and nodal axonal
membrane. The changes in threshold associated with these
electrotonic changes in membrane potential normally have a
similar time course to the changes in membrane potential
and are known as threshold electrotonus. The protocol for
this study used threshold tracking to record the changes in
threshold induced by 100-millisecond polarizing currents, set
to 40% (depolarizing) and 40% (hyperpolarizing) of the
control threshold current. For the sensory fibers, the re-
sponses to polarizing currents 20% of threshold were also
tracked.
The current–threshold relation (see Figs 1C and 2C) was
recorded at the end of 200-millisecond polarizing currents,
which were altered in a ramp fashion from
50% (depolar-
340 Annals of Neurology Vol 57 No 3 March 2005