A review of hazards associated with exposure to low voltages

Dr. Marom Bikson

Department of Biomedical Engineering, City College of New York of the City University of New
York
The Graduate School and University Center of the City University of New York
New York, N.Y
OVERVIEW
This review summarizes peer reviewed papers,
government reports, and regulatory group
recommendations on hazards from electricity. A
goal of this report is to determine a safe voltage
level below which these hazards will not occur; the
review emphasis is, therefore, on extremely-low
voltage (<50 VRMS or 71 VPEAK) exposure. This
report is divided into five main sections dealing with
human exposure. SECTION I addresses the basic
mechanisms by which electric current can affect
biological tissue in a hazardous manner. SECTION II
summarizes experimental research studies involving
application of electric current to human subjects.
SECTION III reviews the epidemiology and case
reports of human electrocution.
SECTION IV
includes a summary of previous electrical safety
standards.
SECTION V includes the review
conclusions for human exposure. An APPENDIX
deals with the electrocution of dogs.
Review Scope
This review and its summary conclusions
relate only to adverse effects of transdermal current
exposure 1 . This review is not concerned with
electric shocks that cause no long-term hazardous
effects (e.g. sensory sensations such as noxious
stimulation and phosphenes). Moreover, this review
does not include injuries that result from humans
being startled by otherwise non-hazardous electrical
current (e.g. falls) or interference with medical
devices. This review only includes scientific reports
which: 1) appeared in scientific journals; 2) include
recommendations by a (inter)nationally recognized
1

When electricity enters the body subdermally, as for

example through two needles inserted into the heart,
voltages as low as 20 V and currents as low as 100 A
can cause fibrillation (Camps et al. 1976). With
electrodes placed directly on the heart, ventricular
fibrillation is usually achieved with voltages of ~0.2 V
and current flow of 80-600 A (Kugelberg 1976;
Webster 1998).

Low Voltage Electrocution

-1-

scientific organization; or 3) were sponsored by a

government agency. This review does not address
resuscitative measures, forensic diagnosis, or electric
safety measures. Lightning strikes, high-voltage
arcs, and electrical fires/explosions are not
considered.
SECTION I: BASIC MECHANISMS BY WHICH
ELECTRIC CURRENTS AFFECT BIOLOGICAL TISSUE

Electrically Excitable Tissue / Burns

The human body will conduct electricity. If
the body makes contact with an electrically
energized surface while simultaneously making
contact with anther surface at a different potential
(or ground) then an electric current will flow
through the body, entering the body at one contact
point, traversing the body, and exiting at the other
contact point. The magnitude of this current will
increase as the voltage difference across the contact
points increases. This section introduces potential
hazards associated with such currents.
Certain tissues in the body have
traditionally been considered most sensitive to
electricity because they normally use bio-electric
signals. Cells in the central and peripheral nervous
system (neurons) use bio-electrical signals to rapidly
process and communicate information. Neurons
regulate the contraction of cardiac cells, diaphragm
muscle cells (inducing lung inspiration), and
peripheral muscle cells (controlling movement).
Cardiac and muscle cells, in turn, also use bioelectric signals to trigger their contraction. These
cells are collectively referred to as electrically
excitable cells (Hille 2001)
Electric stimulation, or electrical
shock, results when a portion of the current
conducted by the body passes through/polarizes
excitable cell membranes. Theoretical and forensic
studies examining the effects of electricity on
biological tissue have thus focused on systems
containing or regulated by excitable cells. For
example, electric shock can lead to activation of

M Bikson

neurons/muscles involved in respiration or cardiac

pacing. Electric shocks can acutely affect cell
function without necessarily damaging these cells.
Electric currents may also heat external and
internal tissue sufficiently to induce structural
damage through electrical burns. Electrical burns
affect human health through actions on both
excitable (e.g. cardiac, nervous) and non-excitable
(e.g. skin, blood vessels) tissues.
Electroporation refers to changes in cell
membrane permeability by electric fields.
Electroporation affects both excitable and nonexcitable cell function and can lead to irreversible
cell damage (Chilbert 1998). The role of irreversible
electroporation in transdermal electrical accidents
remains unclear and is thought to be limited to high
voltages/currents (>200 mA; Reilly 1998).
Electrocution refers to fatalities resulting
directly from lethal current flow through the body.
Conventions, Metrics, and Cellular Models
The theory governing the interaction of
electric current with excitable cells (electric shocks)
has been well characterized by my group (Durand
and Bikson 2001; Bikson et al. 2004) and others
(Reilly 1998; Rattay 1999; McIntyre and Grill 1999).
The consensus of these reports is that the effects of
electric currents can be directly calculated using
information about the detailed cell geometry/
biophysical properties and detailed information
about the electrical potential (induced by current
flow) along this geometry; all with micrometer
Unfortunately, this combined
resolution 2 .
physiological/electric-potential data is not available
for human exposure.
Moreover, theoretical
consideration of the detailed electric potential
induced under various exposure conditions and their
effects on every excitable cell in the body is
intractable (but see Reilly 1998, p334).
Therefore, the approach taken by previous
research reports and regulatory groups has been to
assume a (quasi-) uniform electric field (E in
volts/meter) across the tissue of interest (but see
Reilly and Diamont 2003).3 This assumption may

Specifically, information about the second derivative

of the extracellular potential in space can be related to
the excitation strength, the amount of current moving
across a cell membrane.
3
Note that this assumption ignores a central concept
that it is the second special derivative of voltage (i.e.

Low Voltage Electrocution

-2-

be grossly valid for currents passing across the entire

body.
It significantly simplifies the analysis of
electric current effects because it allows
standardization across disparate experimental studies
where similar uniform fields were used. In
addition,
this
assumption
facilitates
the
establishment of safe exposure levels using a single
number, such as the uniform electric field strength.
Electric field (E in units of V/m) can be
related to current density (J in A m-2) by: E = J
where a homogenous volume resistivity ( in m)
is assumed. Current density can, in turn, be related
to current (I in A) across the entire tissue through
knowledge of the tissue geometry, notably crosssectional area (Acs in m2), and current entry/exit
locations. The voltage (V in V) across the entire
tissue can be theoretically related to current I
through a tissue with a total path resistance R by V =
IR. Finally, for a cylindrical block, total path
resistance R, can be related directly to uniform
resistivity by R = d/ Acs where d is the path
length (in meters).
Stray voltage refers to unintended
electrical potentials between contact points that may
be encountered by humans or animals. Accidental
electrocution can result when stray voltages exceed a
safe threshold voltage level. This report focuses on
determining this safe voltage level as stray voltages
are readily measurable during quality assurance
(Con Edison 2004). The internal electric field will
depend not only on the contact voltage magnitude
but also on contact geometry/tissue properties; safe
voltage levels may thus vary depending on
exposure/subject
conditions.
Additional
consideration must be exercised in considering safe
voltage levels (as opposed to safe current levels);
note that an individual wearing electrically
insulating gear (e.g. rubber gloves) can potentially
be exposed to higher voltages without hazardous
effects; this is because the total resistance including
the insulating gear will be much higher than the
body resistance alone, and the resulting current
flowing through the body will be significantly
reduced. The skin itself (see Body resistance below)
can also provide a significant amount of electrical
insulation.
During exposures of high intensity and long
duration, tissue heating may be a mechanism of
injury. The heating of tissue by electric currents
(burns) is related to the total amount of energy
the first derivative of a non-uniform electric field)
that is critical in determining excitation strength.

M Bikson

delivered to the tissue and the thermal properties of

the tissue. The rate at which energy is delivered by
current flow at any instant is given by its
instantaneous power (p(t) in Joules/s or W) which
can be calculated from p(t) = v(t)i(t) (equivalently
i(t)2R or v(t)2/R) where v(t) is the instantaneous
current and i(t) the instantaneous current. The total
energy (in Joules) over time delivered to a system
can be calculated by integrating this instantaneous
power over time (T).
In the case of constant voltage and current
(dc), total energy can be calculated4 from VIT. In
the case of alternating voltage and current (ac), a
sinusoidal current, through a specific resistive path,
will deliver the same amount of power as a constant
(dc) current with an amplitude 0.707 times the peak
sinusoid amplitude, irrespective of frequency. The
root-mean-square (RMS) value of a sinusoid is
calculated as 0.707 times it peak.
A majority of this energy will be converted
into heat causing a rise in tissue temperature.
However, the body actively regulates temperature,
thus a complete theoretical analysis of electric
current induced tissue heating may be complex; key
bio-heat models have been previously been
developed at The City College of New York (Arkin
et al. 1994). Large temperature rises can lead to cell
death.
It is accepted that all effects of electricity
are aggravated as the amplitude of the
current/voltage is increased.
Consistent with
industry standards5, low voltage refers to voltages
under 1000 VRMS. Very-low voltage is defined
here as between 50-1000 VRMS (71-1414 VPEAK);
many household and commercial electric systems
operate in this range. Extremely-low voltages are
defined here as below 50 VRMS (71 VPEAK).

Or alternatively as Q =J2

energy density in J m-3.

Particular attention is paid in this review to hazards

associated with extremely-low voltages.6
Current temporal waveform and exposure duration
Tissue heating depends on the RMS values
of the current and little or not-at-all on its
waveform/frequency (e.g. the RMS value of a
sinusoid is independent of frequency). However, for
electric stimulation (shocks) the waveform of the
current can have a profound influence on current
efficacy (Bikson et al. 2004). AC (e.g. 60 Hz
sinusoidal) is considered more likely to induce
hazardous electric shocks than dc current (Camps et
al. 1976; Reilly 1998). Dalziel and Lee (1969)
found 10-400 Hz currents most effective in inducing
involuntary hand muscle contraction. DiMaio and
DiMaio (2001) considered 39-150 Hz the most lethal.
Kugelberg (1976) found frequencies between 12-60
Hz most effective in inducing fibrillation of the
human heart. The two competing factors relating to
stimulation frequency efficacy are: 1) as electrical
excitation occurs during the rising or falling phase of
the current flow, increasing frequency increase the
amount of potential excitations per time (e.g.
involuntary muscle contractions); 2) excitable cell
membranes act as low-pass filters (Bikson et al.
2004), higher frequencies are thus less effective.
Unless otherwise stated the results reviewed here
refer to 50/60 Hz sinusoidal waveforms (as are used
in power distribution systems).
The duration of exposure will influence both the
electric shock and burning mechanisms. Increasing
exposure time will result in more energy delivered to
the tissue; if the tissue cannot dissipate this energy,
it will continue to heat up and eventually burn (see
Conventions, Metrics, and Cellular Models). The
long term effects of electric shocks are more
complex (Durand and Bikson 2001). As exposure
6

T where Q is thermal

In the IEEE Standard Dictionary (IEEE Std 100-),

high voltage is defined as follows: for maintenance

of energized power lines, as voltage levels above

1000 VRMS, and for system voltage ratings as a class
of nominal system voltages from 100 to 230 kV RMS.

Medium voltage is defined as a class of nominal

system voltages from 1000 V RMS to 100 kV RMS.. Low
voltage is defined as a class of nominal system
voltages 1000 V RMS or less.

Low Voltage Electrocution

-3-

This review repots voltage/current values in both

PEAK and RMS. In reviewing previous reports,
numbers explicitly given in RMS have been converted
to PEAK. When unspecified in the original report,
numbers are exactly reproduced here. This approach is
consistent with: 1) an appreciation that electric shocks
(rather than heating) and hence peak voltage (rather
than RMS) is most relevant for extremely-low voltage
electrocution; and 2) an attempt to develop a safe
threshold voltage (RMS values are always equal to or
less than peak values). Thus, in this review, unless
otherwise specified, voltage should be read as PEAK.

M Bikson

time increases the chance of an electric shock

precipitating a stochastic reaction increases.
The shorter the exposure duration the
greater the current needed to induce ventricular
fibrillation (see cardiac results below; Camps et al.
1976). DiMaio and DiMaio (2001) suggest that for
either 120 mA or 1.2 A exposure, ventricular
fibrillation could occur in 4 s and 0.1 s respectively.
Dalziel and Lee (1969), extrapolating from animal
studies (APPENDIX), reported that, for exposure
times (T in seconds) between 8.3 ms and 5 s, the
lower 0.5% population-rank current threshold (I0.5f
in mA) for induction of fibrillation is given by:

I 0.5 f = k

(1)

where K is a constant linearly related to victim

weight (see Individual Susceptibility below) and
presumably to contact geometry (see Current Path
below). Dalziel and Lee (1969) concluded that
ventricular fibrillation in a normal adult [>50 kg]
worker is unlikely if the shock intensity is less than
116 T milliamperes [RMS]. In addition, for
longer duration (5 to 30 s) exposures, the threshold
may remain fairly steady. Biegelmeier and Lee
(1980) proposed an inverse relationship between
fibrillation threshold and exposure time (If T-1)
for exposure periods between 0.2 and 2 s; and a
threshold plateau above and below these times (a Z
curve).
The direct effects of electric shocks
terminate once contact with energized source ends.
In some cases, induced pathologies will persist (e.g.
fibrillation).
In other cases, normal function
resumes immediately after termination of
stimulation (see respiratory arrest).
Body resistance/related susceptibility factors
Though the effects of electricity are more
directly standardized by reference to current, for the
reasons discussed above, it is often necessary to
consider contact voltage levels. Ohms law (voltage
= current total path resistance) establishes a simple
relationship between these two factors. Total path
resistance is thus a critical factor in determining
voltage safety levels. As a result of body/electrode
contact geometry and tissue inhomogeneity, the
current density across any given tissue will vary.
Regardless, for any given voltage, decreasing body

Low Voltage Electrocution

-4-

resistance will increase overall current magnitudes

and hence the risk of electrocution.
Generally tissue inhomogeneity is ignored
and resistance is reported as the total current path
resistance, the body resistance, between two
electrodes on the body surface. Note that measured
total body resistance includes the resistance of the
electrodes (see below). Internal body resistance
refers to the total current path resistance excluding
the skin. Experimentally, the skin resistance can be
removed by abrasion. For measuring total body
resistance, common electrode positions include
hand-to-hand and hand-to-foot.
During electrocution the total current path
resistance will be the sum of the body resistance
including the skin, the electrode interface resistance,
and the resistance of any other materials along the
current path such as cloths or soil. High resistance
clothing (e.g. rubber or dry leather gloves, dry shoes)
provides increased protection against electric
hazards; e.g. >1 M (Reilly 1998). As this review
will focus on worst-case scenario exposure,
protective gear and insulating clothing are not
considered further.
There is an expected variation in resistance
across subjects.
In addition, individual body
resistance is a complex function of electrode size,
electrode material, electrode position, skin and its
surface conditions, temperature, and applied voltage
magnitude/frequency (Biegelmeier 1985a; Reilly
1998; Webster 1998). Moreover, skin resistance
will vary with exposure time 7 . The following
section addresses the general range of resistances
that would be relevant during extremely-low voltage
electrocution at frequencies <100 Hz.
In
experimental studies, wet conditions are usually
achieved using conductive NaCl or saline solution.
The presence of moisture from wet clothing,
high humidity, or perspiration will decrease the
electrical resistance of the electrode interface and of
the skin; this will increase current magnitudes and
thus the risk of electrocution. High-humidity is a
common feature in very-low voltage electrocutions
and ubiquitous with extremely-low voltage
electrocution (SECTION III).
The increased
frequency of low-voltage electrocution in summer
7

The total opposition to the flow of ac current is

denoted impedance, which in general is a complex
quantity with a magnitude and phase. For evaluating
electrocution only the magnitude of the body
impedance (rather then the phase) is critical; this
magnitude is a real number.

M Bikson

has been suggested to relate to sweating (Wright and

Davis 1980; Fatovich 1992). Also, the resistance of
the skin is significantly reduced by abrasion/cuts.
The passage of significant current can
reduce skin resistance to negligible values (>200 V,
Biegelmeier 1985a; >240 V, Webster 1998; 100-250
V, Reilly 1998; Camp et al. 1976).
Thus with
higher voltages, skin conditions are considered to
play no significant role (DiMaio and DiMaio 2001).
Gettman (1985) noted that below 12 V skin
resistance would not break down.
Skin can provide a significant source of
insulation. Skin has a surface resistivity (in cm2).
Skin resistance is a function of contact area; larger
contact areas result in reduced effective total skin
resistance. Camp et al. (1976) noted that if the
palms are thickened by manual labor the dry skin
resistance can reach 2000 k . DiMaio and DiMaio
(2001) concluded that for 120 V, dry skin may have
a resistance of 100 k ; dry and calloused skin up to
1000 k ; moist skin 1 k or less; and moist, thin
skin as low as 100 . Wright and Davis (1990)
considered the minimum internal resistance of the
body 500 and minimal hands/feet resistance 1 k ;
dry skin easily reaching 100 k .
Webster (1998) noted dry skin resistance
(one square centimeter area) may range from 15 k
to 1 M ; if the skin is wet or broken this resistance
drops to as low as 1% of the value of dry skin.
Webster (1998) considered the internal resistance of
the body (between any two limbs) 500 , with each
limb contributing ~200 and the trunk ~100 .
Using large-area (>80 cm2) electrodes and
for low voltages, Biegelmeier (1985a) reported a
initial (analogous to internal) body resistance of
781 114
and 637 98
for 25 V and 15 V
respectively (average standard deviation; hand-tohand path). Reviewing data on the distribution of
internal resistance, Biegelmeier (1985a) concluded
that ~50% of total body internal resistance (limb-tolimb) resides in the wrists and ankles (e.g. 25% of
hand-to-hand total path resistance is across each
wrist); the body trunk accounted for <10% of total
limb-to-limb internal resistance. Using a simplified
model, Biegelmeier (1985a) considered 2 hand-to-2
feet and 2 hand-to-trunk internal resistance 50% and
25% of total hand-to-hand resistance, respectively.
As contact area decreased, the total body resistance
increased significantly (for a 0.01 cm2 electrode
resistance was >1000 k at 100V).
Osypka (1963) reported a total-bodyimpedance of ~2 k (hand-to-hand) and 500
(2
hands-to-2 feet) under wet conditions (60 Hz, ~10V).

Low Voltage Electrocution

-5-

Freiberger (1934) reported an average handto-hand and hand-to-foot internal body resistance
(skin ablated) of ~1 k ; these measurements were
conducted on cadavers which could result in an
overestimate of resistance in living persons (Reilly
1998). Freiberger (1934) examined the distribution
of internal resistance; he concluded that ~50% of the
total internal body resistance resides in the wrists
and ankles. The body trunk resistance was <10% of
the total internal body resistance. Freibergers
statistical data was incorporated by the International
Electrotechnical Commission (Table 1).
Table 1. Statistical data for total body impedance as
adopted by the International Electrotechnical
Commission (CEI-IEC 479-1; 1994). Large area
contacts, hand-to-hand, dry skin, ac 50/60 Hz.

Touch voltage (V)

25
50
75
100

Total body impedance ( ) not

exceeded
by
indicated
percentile rank
5%
50%
95%
1,750
1,450
1,250
1,200

3,250
2,625
2,200
1,875

6,100
4,375
3,500
3,200

Using
short-duration
high-voltage
discharges in living people, Taylor (1985) found
similar resistance distributions across the body as
Freiberger (1934) and Biegelmeier (1985a). Taylor
(1985) reported a total body resistance of 470 36
(mean S.D.; hand-to-hand) and 516 55 (left
hand-to-left foot). Taylor (1985) considered 70-100
of these resistances coming from wet skin.
Hart (1985) reported an internal resistance
of 400-500 (hand-to-hand) and 450-500 (handto-foot). He found internal body resistance from
hand-to-forearm was 140
and from finger-toforearm 700-800 .
In the majority of the above studies contact
area was relatively large for the purposes of
reducing electrode resistance (see below) and
reproducing worst-case condition. Using data
from animal models (100 V, dry contact), Prieto et al.
(1985) found total body resistance increased with
decreasing contact area (total path resistance area0.32
) and contact circumference (total path resistance
circumference-0.51).
Statistical impedance data was developed
by Underwriters Laboratories (12 VDC, relatively
large electrodes, wet conditions; Whitaker 1939)

M Bikson

for children (3-15 years; 14-58 kg) and for adults

(18-58 years; 45-95 kg). For children they found a
resistance variation from the 5% to 95% rank of 1.7
to 4.47 k (hand-to-hand) and 0.9 to 2.04 k (two
hand-to-two feet).
For adults they found a
resistance variation from the 5% to 95% rank of 1.28
to 2.45 k (hand-to-hand) and 0.63 to 1.16 k (two
hand-to-two feet). 60 Hz AC resistance values can
be 60-90% of these DC resistance values (Reilly
1998).
Thus the majority of research reports
consider worst-case (large contact area, wet
conditions) total body resistance (limb-to-limb) to be
slightly greater than 500 . Worst-case resistance
across the chest can be less than 100 . Under nonworst-case conditions (e.g. small contact size, dry
skin) total body resistance values quickly increase to
greater than 2 k .
Electrodes / Metal-body interface
Metal in contact with tissue is referred to as
an electrode. In the metal electrode and in
attached electrical circuits, charge is carried by
electrons. In the body, charge is carried by ions,
notably sodium, potassium, and chloride ions.
During the passage of current between the metal
electrode and tissue, as occurs during electrocution,
the central process that occurs at the electrode-body
interface is a transduction of charge carriers from
electrons in the metal electrode to ions in the body.
We have recently reviewed this process (Merrill et al.
2004).
Key to the present review is that this
interface can 1) have a significant resistance and 2)
can alter the waveform of the passing current. The
magnitude of this resistance/distortion is a complex
function of metal electrode composition, the
magnitude of both the applied current and voltage,
and effective electrode contact area (roughness); this
later dependence would contribute to the increased
total body resistance noted with small contacts
above. To my knowledge the effect of the electrodebody interface on waveform has not been considered
in the context of electrocution; changes in current
waveform will affect electric shock thresholds. The
resistance of the interface will act to increase the
resistance of the total current path and thus protect
from electrocution. Under worst-case-conditions
the interface resistance may be assumed to be
negligible.
Mechanisms of injury/death by electrocution

Low Voltage Electrocution

-6-

Electrocution by extremely-low voltages

(<71 VPEAK) is significantly less common then by
very-low (71-1414 VPEAK) or higher voltages
(SECTION III). Literature on the pathophysiology of
electrical injury is therefore dominated by higher
voltages. Unless otherwise specified, the comments
in the following section do not necessarily apply to
extremely-low voltage exposure.
The cardiac results of electrical injury may
be fatal (Leibovici et al. 1995). Low-voltage
exposure can induce ventricular fibrillation. High
voltage/current electrical exposure can lead to
aystole, which renders the heart electromechanically
silent. These heart pathologies can be elicited by
stimulation of cardiac muscle cells or pacemaker
neurons. Both pathologies may persist even after
termination of current flow. Because normal blood
pumping stops, if the victim is not actively
resuscitated (e.g. defibrillated), unconsciousness can
follow in 10-15 s (Wright and Davis 1980),
irreversible damage can occur within 3 minutes
(Geddes and Baker 1989), and death in 5-10 minutes
(Dalziel and Lee 1969; Wright and Davis 1980). The
brain and heart are particularly sensitive to anoxia
(Cabanes 1985). The mechanisms of ventricular
fibrillation have been reviewed elsewhere (Bridges
et al. 1985).
Table 2 summarizes the currents considered
sufficient to disrupt cardiac function. The threshold
for inducing fibrillation is dependent on exposure
duration (see Exposure duration above) and body
weight (see Individual susceptibility below).
Grouping all (high and low voltage) electrocutions,
cardiac fibrillation is considered the most common
cause of death (Fatovich 1992; Webster 1998).
Blood vessels are also sensitive to electric currents
(Leibovici et al. 1995).
Electric currents can induce respiratory
arrest though both tetanic paralysis of the
respiratory muscles and damage to the respiratory
control centers of the brain. In both cases death is
asphyxial with the heart continuing to beat until
death occurs.
Irreversible respiratory system
damage is considered rare (Camps et al. 1976);
indeed the passage of several hundred milliamps
through the brain during therapeutic electroconvulsive therapy rarely damages the respiratory
control centers (Devanand et al. 1994). Tetanic
paralysis is considered more common and requires
continuous contact with the energized source
(Wright and Davis 1980; DiMaio and DiMaio 2001).
In contrast to cardiac results, the effects of current

M Bikson

on respiration have not been extensively studied.

Dalziel and Lee (1969) reported that 18 to 22 or
more milliamperes [RMS], flowing across the chest
stopped breathing during the period the current
flowednormal
respiration
resumed
upon
interruption of the current, and no adverse aftereffects were produced. Cerebral injury, as result of
hypoxia, can occur if the duration is more than 3 to 4
minutes (Cabones 1985). Table 2 summarizes the
currents considered sufficient to induce tetanic
paralysis of respiratory muscles.
Electric shock can induce violent muscle
contraction and hemorrhage of muscle fibers
(Leibovici et al. 1995; Karger et al. 2002). This can
lead to dislocation of joints, fractures, and
fatal/immobilizing falls. DiMaio and DiMaio (2001)
summarize cases of fractures resulting from 50/60
Hz currents ranging from 110-440 V. In addition,
violent muscle contraction can damage blood vessels;
this can aggravate, in feed-forward manner, the
direct effects of electric fields on blood vessels
(Leibovici et al. 1995). Non-fatal hand muscle
contraction is discussed below (let-go current). In
the case of contact with an energized source on the
floor, falls resulting from leg muscle contraction can
subsequently expose a victim to currents across the
chest (see Current Path).
Electric shock can be hazardous through
actions on the nervous system (Leibovici et al.
1995).
Stimulation of the vagus nerve
(parasympathetic system) can induce slowing of the
heart. Stimulation of sympathetic nerve fibers will
have the opposite effect and can trigger
vasoconstriction. Though occurring at relatively low
current thresholds, the above effects are not
necessarily fatal; autonomic influences may affect
ventricular fibrillation threshold (see Individual
susceptibility). Respiratory arrest (see above) can
result from (reversible) electrical stimulation of CNS
breathing centers or the phrenic/thoracic nerves
which produces tetanic contraction of the diaphragm
(Geddes and Baker 1989). Currents flowing through
the CNS can induce cerebral edema, convulsions,
coma, hysteria, amnesia, auditory and visual
dysfunction, and transient loss of consciousness
(Chilbert 1998); it should be emphasized that these
effect has not been demonstrated for extremely-low
voltages. With high currents transient or permanent
neuropathy can occur, especially in limbs that were
in the currents path. Damage can results from
neuronal electric shock, burns, or secondary to
violent muscle contraction.

Low Voltage Electrocution

-7-

Eyes can be damaged by electric shocks

when the voltage is greater then 200 volts (Leibovici
et al. 1995; Boozalis et al. 1991).
It has been suggested that all high-voltage
but only some low-voltage electrocutions result in
visible burns on the body (Wright and Davis 1980;
Fatovich 1992; DHHS 1998; DiMaio and DiMaio
2001). High-voltage (high-current) burns can result
in limb amputation and death (Chilbert 1998). In
deaths resulting from internal burning, external
burns are expected; the exception being when the
victims were under water or very wet (Karger et al.
2002; Leibovivi et al. 1995). However, very-low
voltage electrocutions, perhaps relating to the
necessary exposure time, are generally associated
with burns (SECTION III); electrical burns can also
occur postmortem.
For low-voltages, local skin burns, usually
limited to points of current entry and exit, are
generally not the vital reaction (the direct cause of
death). Wright and Davis (1980) calculated that the
minimum current necessary to induce cardiac
fibrillation (see above) would induce a peak
temperature rise of only 0.145C. Moritz and
Henriques (1947) reported first-degree burns require
exposure of the skin to 50C for 20 s. Chilbert (1998)
summarized: cutaneous burns occur when the
temperature is elevated for a sufficient length of
time: 45C requires more than 3h, 51C requires less
than 4 min, and 70C requires less than 1s for
injury. Reilly (1998) noted that for currents
passing across a limb, strong muscle contraction and
severe pain due to electrical activation of nerves
would occur at current levels well below those
causing burning. Leibovici et al. (1995) suggested
thermal injuries were the most harmful mechanism
when grouping both low- and high-voltage
exposures. However, the role (if any) of internal
organ heating during extremely-low voltage contact
remains unclear.
Current Path
The body acts as a volume conductor. The
points of current entry and exit are important
because 1) the current density will be highest nearest
these points and 2) the direction of current flow
(electric field) along excitable tissue will affect
electric shock efficacy (Reilly 1998).
When current is applied at two points on the
surface of the body only a small fraction of the total
current flows through the heart (Webster 1998).
Freiberger (1934) reported that for hand-to-feet and

M Bikson

foot-to-foot electrode contacts less then 8.5% and

0.4% (respectively) of the net current would travel
through the heart.
Leibovici et al. (1995) reported that current
passing through the thorax is associated with 60% of
electrocutions, whereas for current passing from leg
to leg it is 20%. These numbers do not address the
over-all (including not fatal) chances of exposure at
each geometry nor do they distinguish across
high/low voltage exposure levels. Leibovici et al.
(1995) note that though current density will be
higher across limbs (due to smaller cross sectional
area), the presence of vital organs in the torso
accounts for lethality of trans-thoracic currents (see
Mechanisms of injury).
Camps et al. (1976) concluded that for
ventricular fibrillation the most dangerous current
path is left arm-to-opposite leg; from arm-to-arm
being 60% less lethal. Bailey et al. (2001) found a
majority of victims died from current flow from
upper-to-lower extremities. In contrast, Alexander
(1941) noticed that more victims die from current
flow from upper-to-upper extremities. As noted
above, these findings would be more relevant if the
prevalence of the current exposure pathway,
including survivors, was known. The role of current
path has been examined systematically in dogs
(APPENDIX).
The resistance of the current path is highly
dependent on contact location (see Body Resistance).
A majority of extremely-low voltage electrocution
cases (SECTION III) involved electrode contacts to
the chest.
Demographics patterns, Individual susceptibility
Fatal
electrocutions,
particularly
occupational, occur predominantly among 20-34
year old males (DHHS 1998; Taylor et al. 2002;
Bailey et al. 2001; Fatovich 1992); this reflects the
prevalence of this demographic in trades with
greater exposure to electrical sources and a generally
higher exposure to electrical equipment and
machinery.
It has been suggested that pre-existing
cardiac pathology can increase risk of death during
or after electric shocks (Camps et al. 1976; Griffin
1985; Bailey 2001). As with other hazards, physical
or mental fatigue or awareness will increase the
chance of exposure; however it remains unclear if
physical fatigue in humans increases susceptibility
to electrocution. Experiments with dogs (APPENDIX)
have shown that during common abnormal

Low Voltage Electrocution

-8-

physiologic conditions there is a reduced current

threshold for producing ventricular fibrillation; in
particular adrenergic stimulation (flight or fight
reaction) and acidosis. In humans, physical exertion
will produce acidosis while aggravating electric
shocks can trigger adrenergic stimulation (Wright
and Davis 1980). Indeed, extremely-low voltage
electrocutions are often associated with physical
exertion and require prolonged contact (SECTION III).
Reviewing previous studies, Griffin (1985)
concluded physical stress, but not exercise (in the
absence of ischemia) reduced fibrillation threshold,
while alcohol may increase fibrillation threshold.
Griffin
(1985)
notes
that
vagal
or
parasympathomimetic activity tends to promote and
increased ventricular fibrillation threshold opposing
the decreased fibrillation threshold resulting from
sympathetic activity.
Some studies using animals have shown
that average fibrillation threshold increases with
body weight (Webster 1998; Reilly 1998). Dalziel
and Lee (1969), extrapolating from animal studies
(APPENDIX), concluded that, the threshold for
cardiac fibrillation is linearly related to weight;
equation (1). Using data from 104 animals of
several species (rabbits, monkeys, dogs, goats,
ponies), Geddes et al. (1973) found the threshold
current for fibrillation, at 5 s exposures, varies
almost as the square root of body weight (W in kg):

I f = bW a mARMS

(2)

where b and a are parameters based on electrodeposition only (e.g. b=29.7, a=0.51 for right forelimbto-left hindlimb path).
Increased mass/volume would be expected
to decrease current density across any given tissue
Based on the assumption that
(SECTION I).
fibrillation threshold is a function of current density
and assuming a fixed shape/organ proportions across
different sizes and weights, Bridges (1985) proposed
a 2/3 power weight relationship:

I f W

(3)

Dalziels linear rule (1), Geddes square

root rule (2), and Bridges 2/3 power rule (3) are in
reasonable agreement over a fairly wide range of
body weights. Some researchers have argued that
these threshold-weight rules are consistent across
experimental species, thus suggesting they may be
quantitatively extendable to man (Dalziel and Lee

M Bikson

1969).
Other groups raise concerns that
experimental animal models have radically different
proportions
between
anatomical
features,
invalidating extension of these rules to human. Even
a qualitative relationship between fibrillation
threshold and weight has been challenged
(Biegelmeier 1985b) based on the concept that the

fraction of current intersecting the heart will not

vary with size/weight. It is generally accepted that
fibrillation data from dogs (APPENDIX) may be used
to establish lower safety limits for man.
See also comments on variations in Body
resistance above.

Table 2. Estimated effects of transdermal 60 Hz AC currents through the human body (in
mAPEAK, limb contact) in selected reports, reviews, and books
DHHS
(1998)
Barely perceptible
Painful Sensation
Tetanic muscle
contraction / letgo current
Paralysis of
respiratory muscles
Ventricular
fibrillation
(exposure time)
Cardiac standstill /
internal organ
damage

Leibovici
et al.
(1995)
5
15

22@
28@
100

Bridges et
al. (1985)
0.5-2

Camps
et al.
(1976)

DiMaio and
DiMaio
(2001)
1

Wright and
Davis
(1980)
1

Webster
(1998)

8-13@

15-17

22@

8-31@

8-25

50

28@

25-31@

25-31@

75-100
(>5 s)

120 (0.1 s)

75-400

33 (5 s)*
807 (8.3ms )

1000

>2000

10006000

8-28@
20-40

(debated)

30-50

2000

>70 (0.1s)
40-100
( )

70 (5 s)

Reilly
(1998)

0.7-2

@
*

(Presumably) based on work by Dalziel and colleagues; adjusted for Peak.

Dalziel and Lee (1969). Lower 0.5% estimate for a small child. Values more than double for adults.

SECTION II: EXPERIMENTAL RESEARCH STUDIES

INVOLVING APPLICATION OF ELECTRIC CURRENT
TO HUMAN SUBJECTS.
Application of currents across the hands / Let-go
current
With experimental subjects, application of
current either hand-to-hand or hand-to-foot can lead
to: 1) direct activation of sensory fibers and hence
discomfort and pain, subjects have a pain tolerance
level at which they let-go; 2) stimulation of
involuntary contraction of both fore-arm flexor and
extensors, when the stronger flexors dominate the
subject may not be able to release, let-go, a
grasped object. In studies conducted from the
1930s to 1950s, both criteria have been used to
establish let-go thresholds.
The former is
dependant largely subject motivation and is not
directly relevant for accidental electrocution.
Because subjects may become attached to an

Low Voltage Electrocution

-9-

energized source, the involuntary contraction letgo threshold (either release grip or rotate handle)
has been of interest in establishing electricity safety
levels. Women have a lower average let-go
threshold then men; this could relate to both physical
and motivational (physiological) factors (Dalziel and
Lee 1969).
Gilbert (1939) used the release-grip
endpoint in determining an average let go current of
21 mA.
Whitaker (1939) using the release-grip
endpoint current found a range of 8.4 mA to 14 mA,
average 11 mA. Thompson using a rotate-handle
endpoint reported an average let-go current of 11.7
mA, maximum 28 mA. In the commonly referenced
initial report of Dalziel (1938), using the release-grip
criteria, the average endpoint was 17.7 mA,
maximum 25 mA (male subjects). These reports
were reviewed by Reilly (1998; Table 2). Dalziel
and Lee (1969, 1972) summarized results from 124
males and 28 females; the average let-go currents
were 22.3 mAPEAK male and 14.8 mAPEAK female

M Bikson

while the lower 0.5 percentile values were 12.7

mAPEAK male and 8.5 mAPEAK female. Thus across
studies 8.5 mAPEAK appears a safe let-go threshold.
Studies using pain criteria are relevant for
electrocution from the perspective that subjects
never experienced chronic adverse effects from the
repeating testing. Grayson (1931) used current in
range of 11.2 mA.
Dalziel and colleagues used
currents as high as 31 mA (Dalziel 1938; Dalziel and
Lee 1969).
The usual let-go test procedure had the
subject firmly grip an energized conductor (often
moist with saline solution). In most cases only
subjects in good physical condition (circulatory,
respiratory systems) were accepted. In applying
these values to electrical safety standards several
important points must be considered 1) the
conditions of contact with a moist hand wrapped
around a handle/wire represents a worse case
scenario; 2) even at the let-go current the subject
may have been able to disengage their hand by
moving other body parts; 3) physiological factors
profoundly effect let-go current levels (Dalziel and
Lee 1969); most importantly, 4) in none of the above
studies were repeated current exposures reported to
result in acute or chronic health problems. Thus the
let-go current does not imply a hazardous level of
current for healthy subjects 8 . Nonetheless the
prospect of a subject becoming attached to an
energized conductor is of concern and this value can
be considered in that sense. Finally, current applied
across the chest may have different thresholds for
safety and mechanism of damage. The let-go
current does provide a rough indicator of the current
levels necessary to tetanize muscle.
Gettman (1985), summarizing Underwriters
Laboratories safety criteria for electric fences,
indicated that acceptable output current decreases
with increasing exposure duration. For 5 ms
exposure duration, currents up to 77 mA were
considered safe; whereas for 200 ms exposure,
currents up to only 7 mA were considered safe.
Gettman (1985) considered a current interruption of
750 ms sufficient to allow individuals to break
contact with an energized wire.
Application of currents to measure resistance
8

Indeed in one anecdotal report a worker became

attached to an energized overhanging wire which acted
to break his fall and thus potentially prevent injury.
Once assistance arrived and the worker was freed from
the wire he suffered no severe chronic effects.

Low Voltage Electrocution

- 10 -

Numerous studies have repeatedly applied

voltages to subjects to measure the resistive
properties of the body (see also above). While these
studies were not explicitly intended to evaluate the
risks of electric shocks, they provide data regarding
shocks that are non-hazardous. Indeed if shocks
were considered reasonably likely to induce any
injury, the studies would not have been conducted.
Underwriter Laboratories has routinely
applied up to 17 V/21 mA (hand-to-hand, hand-tofoot) to adults and children (range 3 to 58 years;
Whitaker 1939). Osypka (1963) applied voltages of
~10 V. Biegelmeier (1985a) used voltages up to 150
V (30 V routinely) for 100 ms.
SECTION III: REPORTS OF ACCIDENTAL HUMAN
ELECTROCUTION.
The
following
section
summarizes
published reports and reviews of electrocution death.
Only studies in which voltage levels were noted are
reviewed here. It should also be emphasize that
particular focus was placed in this review on
extremely-low voltage electrocutions; however these
represent a minority of electrocution deaths.
Based on the National Traumatic
Occupational Fatalities (NTOF) surveillance
systems of the U.S. Department of Health and
Human Services (DHHS) National Institute of
Occupational Safety and Health (NIOSH), 5,348
workers died of electrocution from 1980 to 1992
(DHHS 98-131). The NIOSH Fatality Assessment
and Control Evaluation (FACE) program analyzed
211 electrocution fatalities from 1982 to 1994. The
FACE reports concluded that 147 fatalities resulted
from voltages over 600 V and 25 fatalities resulted
from voltages between 110-120; no electrocution
cases from voltages below 110 V were reported
(DHHS 98-131). The FACE report concludes that
most of these electrocutions could have been
prevented through compliance of existing OSHA,
NEC, and NESC regulations; and/or use of adequate
personal protective equipment (PPE). Table 2
includes a summary of the DHHS (1998) estimates
of the effects of 60 Hz ac currents.
Peng and Shikui (1995) reviewed cases of
low-voltage electrocutions in China. Though they
state that it is relatively rare to be electrocuted by
voltages lower than 100 V, specific statistics or the
methods by which they identified cases were not
presented. Peng and Shikui (1995) presented 7 cases
of electrocution by AC or DC voltages ranging from

M Bikson

25-85 Volts. In all cases, the contact site was on or

near the chest, the contact time was long, and skin
burns were observed. In addition the authors note
that all victims were working in an enclosed, high
humidity and high temperature environment which
would 1) increases susceptibility to electric shock
through decreased skin resistance; 2) decrease
reaction time and ability to disengage from the
voltage source; 3) increase the chance of heatstroke
and unconsciousness as a result of exertion/fatigue.
The victims were otherwise healthy 20-41 year old
males. The authors note that autopsies showed
congestion of internal organs and some focal
hemorrhage. These changes comforted asphyxial
death ventricular fibrillation might also exist.
Bailey et al. (2001) reviewed 124
electrocution deaths in Quebec between 1987 and
1992; 25 fatalities occurred at 240-120 V (RMS), the
lowest voltage range reported.
Transthoracic
currents (59%) and wet extremities (50%) were
more common in the 240-120 V range than for
fatalities resulting from higher voltage (>240 V)
electrocutions, while burns where less common for
low voltage electrocution.
Karger et al. (2002) re-examined 36 cases of
German electrocution fatalities from 1972-1997.
Twelve of these fatalities (32%) occurred in the
voltage range of 65-1000 V, the lowest voltage
range reported. In this lower range, burns were less
common than in the higher (>1000 V) range.
Fatovich
(1992)
summarized
104
electrocution fatalities in Western Australia from
1976 to 1990. He reported that 88% of victims were
exposed to voltages less then 1000 volts.
Wright and Davis (1980) investigated 220
deaths by electrocution in Dade County, Florida.
They report that 108 deaths resulted from voltages
below 1000 V.
It is important to note that in all the above
reports, voltages were measured at some period after
the electrical accident and with loads/geometries
potentially different than during the electrocution.
In some cases electrocution may be overlooked as a
cause of death (Wright and Davis 1980) while in
others it may be incorrectly reported as the cause of
death. Thus while case reports of electrocution
remain a decisive indicator of safe voltage
thresholds, care must be taken in accepting their
quantitative findings.

Low Voltage Electrocution

- 11 -

SECTION IV STANDARDS
There is no single internationally enforced
electrical safety code 9 . In the United States, the
Department of Labor Occupational Safety and
Health Administration (OSHA) standards are
federally enforced. In addition, individual states and
municipal authorities have adopted specific codes,
notably the National Fire Protection Association
(NFPA) National Electrical Code (NEC).
Electrical safety codes often do not specify
a safe voltage level or fail to provide scientific
justification for specified safe levels. In the latter
case it is not possible to tell if specified safe level
incorporate an (arbitrary) safety factor. A safety
factor refers to a reduction in specified exposure
levels below levels scientifically established to cause
harm. In addition, it is not always specified if
standards are lowered to protect from non-hazardous
(e.g. startling) electric shocks. In this section,
underlined voltages result from codes which provide
(minimal) scientific justification.
Standards for medical devices, which may
be used under open heart conditions1, are
peculiarly stringent 10 ; these standards are not
included in this review.
The 2004 OSHA Electrical Standard;
Proposed Rule (29 CFR Part 1910) does not
explicitly define a safe voltage level but does
specify: except as elsewhere required or permitted
by this standard, live parts of electric equipment
operating at 50 volts or more shall be guarded
against accidental contact by use of approved
cabinets or other forms of approved enclosures.
OSHA specifications traditionally draw heavily
from the NEC (NFPA 70) and the NFPA Standard
for Electrical Safety in the Workplace (NFPA 70E).
The 2002 NEC (NFPA 70) applied reduced
safety standards for exposure to extremely-low
voltages (all uninsulated parts <71 VPEAK / <50
VRMS). The 2002 National Electrical Code specifies
that under non-hazardous conditions live parts of
generators operated at more than 50 volts [RMS] to
9

In this review standard and code are used

equivalently.
10
The Association for the Advancement of Medical
Instrumentation developed an American National
Standard on Safe Current Limits for Electromedical
Apparatus (ANSI-AAMI ES1-1993) which limits
leakage current to less than 1 mA. The International
Electrochemical Commission (IEC) 601-1 standard
allows a patient auxiliary current up to 100 A.

M Bikson

ground shall not be exposed to accidental contact

when accessible to unqualified persons. (445.14)
and that where wet contact (immersion not included)
is likely to occurVmax shall not be greater than 15
volts [RMS].
The 2004 NFPA 70E standard,
which is compatible with corresponding provisions
of the NEC, summarizes that: energized parts that
operate at less than 50 volts [RMS] to ground shall
not be required to be deenergized if there will be no
increased exposure to electrical burns or to
explosions due to electric arcs.
The U.S. Consumer Product Safety
Commission (CPSC) Electrical Safety In and
Around Pools, Hot Tubs and Spas suggests ground
fault circuit interrupters are necessary for
underwater lighting circuits operating at more than
15 V. The CPSC Requirement for electrically
operated toys or other electrically operated articles
for use by children (CPSC 1505) specifies that a
potential of more than 30 volt r.m.s. (42.4 volts peak)
shall not exist between any exposed live part in a toy
and any other part or ground. Exceptions are
provided for protected lamps.
The
International
Electrotechnical
Commission (IEC) has issued several reports on
electrical safety. The IEC Electrical installations of
buildings report (IEC 60634-4-41:2001) specifies
that for unearthed circuits if the nominal voltage
does not exceed 25 V a.c. r.m.s. [35 VPEAK] or 60 V
ripple-free d.c., protection against direct contact is
generally unnecessary; however, it may be necessary
under certain conditions of external influences
(under consideration). For earthed circuits the IEC
considers protection unnecessary when nominal
voltage does not exceed 25 V a.c. r.m.s or 60 V
ripple-free d.c., when the equipment is normally
used in dry locations only and large-area contact of
live parts with the human body is not expected; 6 V
a.c. r.m.s. [8.5 VPEAK] or 15 V ripple-free d.c. in all
other cases.
The IEC report also considered
measures limiting current flow to 30 mA or
measures limiting exposure to voltages of 50 V to 5
s protective. No scientific justification is provided
in the IEC 60634 report for any of the above values.
The IEC Extra-low voltage-limit values
report (CEI-IEC 1201:1992) specifies 22.4 V nonhazardous (worst case conditions of large contact
area, negligible skin and earth resistance); the report
notes higher voltages are not necessarily
hazardous. The repots also specifies that for a
non-grippable part with contact area less 1 cm2 the
limit increases to 92.4 Volts.

Low Voltage Electrocution

- 12 -

The most comprehensive review of

electrical safety to date by the IEC, Effects of
current on human beings and livestock (CEI-IEC
479-1; 1994) includes an analysis on effective body
resistance (Table 1).
The report notes that
conductive solutions decrease the impedance
considerably down to half the values measured
under dry conditions [e.g. 875 for 5% rank at 25
V].impedance for children is expected to be
somewhat higher.
The threshold values for
ventricular fibrillation, the main cause of death by
electrical shock, were 40 mA for >3s exposure, 50
mA at 1 s exposure, and 500 mA for 0.1 s exposure
(left hand-to-foot/feet). An identical threshold was
set for breathing arrest. With one electrode on the
chest the threshold for ventricular fibrillation by
only a rough estimation reduces to 60% of the
above values. The CEI-IEC 479-1 report did not
explicitly specify a safe voltage level; however, a
current of 40 mA across 875
corresponds to a
voltage of 35 V (left hand-to-foot path). A new IEC
report directly addressing touch voltages is
expected (IEC-61201).
The 1988 European Organization for
Nuclear Research, (CERN) Dangers due to
electricity safety Instruction (IS-28) is essentially
based on IEC publication 479-1 (see above)
including statistical resistance values (Table 1). In
addition the IS-28 report specifies a total body
resistance of >650 under moist/wet conditions and
>325 for immersed skin. Presumably based on
minimal let-go thresholds, the IS-28 report
specifies as a rough guide to complete safety, the
current limit should be considered as 10 mA for
<20 ms.
The IEEE National Electrical Safety Code
(C2-2002) explicitly does not specify a minimum
approach distance for exposed parts energized below
71 V (Table 441-1). The IEEE Guide for safety in
AC substation grounding (80-1986) incorporates a
voltage-independent body resistance of 1000 and
further considers the insulating effect of soil
between the human body and contact points; the 801986 standard uses ventricular fibrillation to
establish current limitations.
Lee and Meliopoulos (1999) analyzed the
IEEE 80-1986 standard and the IEC-479 standard
with the addition of soil resistance. For their worstcase-conditions (soil resistivity 10 cm, 0.5 s shock
duration, hand-to-2 feet contact) they calculate
permissible voltages of 166 V and 89 V for the IEEE
80-1986 and the IEC-479 standard (with soil
resistance), respectively.

M Bikson

SECTION V REVIEW SUMMARY

Conclusions

In practical situations, uninsulated contacts

with an energized source will most likely involve
two or more limbs.
Based on the reports
summarized above, assuming two limbs make largesurface-area wet-conductive contact, minimal skin
resistance (but no abrasions or immersion), and
negligible electrode interface resistance, total current
path resistance can be as low as 550 .
Tetanic peripheral muscle contraction (letgo current) does not directly lead to death and is not
clearly linked with low-voltage electrocution. Letgo currents are integrated by some regulatory
agencies because exposure to higher hazardous
current, in combination with loss of mobility due to
muscle contraction can be fatal.
Based on the reports summarized above,
respiratory paralysis can lead to death but requires
several minutes of contact; currents as low as 30
mAPEAK across the chest may induce paralysis.
Assuming a low body resistance of 550
this
corresponds to a voltage of 16.5 VPEAK; this is a
theoretical value based on several worst case
resistance assumptions and limited current threshold
data11. Based on the case studies summarized above
the single lowest voltage reported to cause
transdermal electrocution in an adult is 25 V. Note
that this value is significantly above the threshold
for perception/discomfort (1-5 mA; Geddes and
Baker 1989; Webster 1998; Reilly 1998; Table 2)
while well below the current levels causing
unconsciousness (>300 mA); electrocution by
respiratory paralysis may thus require physical
immobilization not directly related to current flow
(e.g. fall).
Based on the reports reviewed above, for
less than 1 min duration electrical contact, currents
>40 mA may be necessary to cause ventricular
fibrillation, corresponding to theoretical value of
27.5 VPEAK (assuming a 550 body resistance). For
a less than 1 s exposure these values increase to
>100 mAPEAK and 55 VPEAK. In the case reports
summarized above, there were no instances of
accidental electrocution from short-term transdermal
exposure to voltages below 50 VPEAK. Indeed, such
exposures were used routinely during experimental
studies on human subjects.

Only a small percentage of electric shocks

have hazardous effects12. Less than 1000 deaths per
year in the United States result from (non-lightning)
electrical accidents (Chilbert 1998; Bailey 2001; but
see Wright and Davis 1980); one-third of these
deaths are caused by voltages below 1414 VPEAK
(1000 VRMS).
Extrapolating from the reports
summarized above, <200 deaths per year in the
United States result from voltages below 250 V.
Reports of electrocution below 50 V are rare.
Calculation of general voltage safety values
from fundamental (cellular) principles is currently
intractable (SECTION I). The electrocution reports
summarized here (SECTION III) indicate 25 VPEAK
(17.8 VRMS) is the lowest lethal voltage; an extreme
combination of sensitivity factors is required for
extremely-low voltage electrocution.
The
combination of both worst-case experimental
resistance and worst-case experimental current
values (SECTION I/II, above) provides theoretical
values below this threshold; this approach is based
on limited data and speculative exposure conditions.
A 25 VPEAK safety threshold is conservative relative
to most electrical codes and standards (SECTION IV);
OSHA, NFPA 70E, and NEC, all have more
stringent requirements for voltages above 71 VPEAK
(50 VRMS).
In the unlikely event a victims exposed
chest contacts an energized source, the safe voltage
threshold would be reduced. It must be emphasized
that significant unknowns remain about the effects
of electricity on the body and further experimental
studies are necessary to better establish under what
conditions extremely-low voltages can harm humans.
Specific areas of concern include determining
current threshold for respiratory arrest, risks under
immersion conditions, and the need for separate
resistance/current thresholds for children.

12

11

The 30 mA value, though cited in several reviews, is

apparently based solely on work by Dalziel et al.

Low Voltage Electrocution

- 13 -

Cases of non-fatal electric shock are often not

reported. There are several reports of legal
electrocution incidents where voltages >1400 V
applied for >17 s did not result in death
(http://hypertextbook.com/facts/NancyRyan.shtml).

M Bikson

APPENDIX: ELECTROCUTION OF DOGS.

Research on the effects of electrical currents
on canines has been conducted towards better
understanding and treating human exposure. In this
context, some studies with canines have been
discussed above. This section deals specifically
with the risks of dog electrocution.
Similar
mechanisms are thought to result in electrocution in
dogs as in humans (see Mechanisms of injury above).
As in humans, cardiac fibrillation has been the most
extensively studied.
Dalziel and Lee (1969) consolidated reports
by Ferris (1936), Kouwenhoven (1959), and Kiselev
(1963) examining fibrillation thresholds in dogs.
Dalziel and Less (1969) noted the inverse
relationship, across studies, between current
threshold and the square root of exposure time,
equation (1) above. The lower 0.5-percentile-rank
for fibrillation thresholds for 5 s, 3 s, and 8.3 ms
exposures were 35 mA, 45 mA, and 910 mA
respectively (5 to 27 Kg dogs); smaller dogs had on
average lower fibrillation thresholds. In the above
tests, electrodes were attached to the right front foot
and the opposite rear limb; Kouwenhoven et al.
(1932) considered current intensity more important
in this configuration than from upper extremity-toupper extremity.
Geddes and Baker conducted a series of
experiments examining the effects of electric
currents on dogs (summarized in Geddes and Baker
1989). They note: A consistent sequence of events
occurred with each animal as the current was
increased. The first was a strong contraction of
skeletal muscles; this was followed by an arrest of
spontaneous respiratory movements, vagal slowing
of the heart, and initially in most animals there was
evacuation of the bladder and bowel. Finally,
ventricular fibrillation occurred. Geddes et al.
(1973) found the current needed for fibrillation using
the upper extremity-to-upper extremity path is about
three times greater than for when current is applied
between the fore and hind limbs. Geddes et al.
(1973; 1989) also found similar thresholds and
duration relationship as summarized by Dalziel and
Lee (1969) with smaller animals having a lower
average threshold. Geddes and Baker (1989) note
that the value of 60-Hz current required to initiate
fibrillation in a 70-kg animal (which approximate
the weight of an adult man) is 215 mA (flowing for
5 s; left arm-left leg). Geddes and Baker (1989) also
examined the effect of current frequency on
fibrillation threshold. The threshold in the region of

Low Voltage Electrocution

- 14 -

10-100 Hz was almost independent of frequency;

above 100 Hz the current required to initiate
fibrillation is markedly higher.
Kugelberg (1976) found frequencies
between 12-60 Hz most effective in inducing
fibrillation of the canine heart; this is within the
range used for power distribution.
Adrenergic stimulation (Han et al. 1964)
and acidosis (Gerst et al. 1966) will reduce the
current threshold for producing ventricular
fibrillation; these conditions would be expected
during physical excursion or (electric shock induced)
stress.
Practically, dogs may be prone to accidental
electrocution as a result of: 1) lack of insulation,
such as shoes; 2) lack of potential hazard avoidance;
3) inability to understand/separate from electrical
hazard once encountered; 4) a higher incidence of
>2 limbs in the current path which would reduce
path resistance.
Cases of accidental canine
electrocution have been reported but the frequency
of these events remains unknown. As in man, once
initiated by a brief electric shock, ventricular
fibrillation rarely stops in dogs and can lead to death
in several minutes. As in man, respiratory arrest has
a lower current threshold but would require constant
exposure for several minutes.
ACKNOWLEDGMENTS
The author thanks Dr. William H. Bailey
and Dr. Duncan Glover at Exponent for comments
on the manuscript. Dr. Bailey is a Principal Scientist,
managing Exponent'
s Health/Epidemiology staff in
the New York, New York, office. Dr. Bailey
specializes in the application of state-of-the-art
assessment methods to environmental health and
impact issues. Dr. Bailey is a visiting scientist at the
Cornell University Medical College and member of
a working group that advises a committee of the
World Health Organization on risk assessment,
perception, and communication. He was formerly
Head of the Laboratory of Neuropharmacology and
Environmental Toxicology at the New York State
Institute for Basic Research, Staten Island, New
York, and an Assistant Professor and NIH
postdoctoral fellow in Neurochemistry at The
Rockefeller University in New York.
Dr. Duncan Glover is a Principal Engineer
specializing in issues pertaining to electrical
engineering, particularly as they relate to failure
analysis of electrical systems, subsystems, and
components, electric power systems, generation,

M Bikson

transmission, and distribution, and power system

planning. Prior to joining Exponent, Dr. Glover was
an Associate Professor in the Electrical and
Computer Engineering Department of Northeastern
University and also held several engineering
positions with companies that include the
International
Engineering
Company,
Commonwealth Associates, Inc., and American
Electric Power Service Corporation. Dr. Glover
holds a Ph.D. in Electrical Engineering from the
Massachusetts Institute of Technology and is a
registered professional electrical engineer in the
state of Massachusetts.
__________________________________________
Comments on the reviewer
Marom Bikson received his B.S. in
Biomedical Engineering (Electrical Engineering
Concentration) from Johns Hopkins University. Dr.
Bikson earned a Ph.D. in Biomedical Engineering
(Neural Engineering) from Case Western Reserve
University where he was a Whitaker and NIH fellow.
Dr. Bikson conducted Post-Doctoral research at the
University of Birmingham U.K. Dr. Bikson is
currently faculty at Department of Biomedical
Engineering, City College of New York of the City
University of New York and faculty of the Graduate
School of the City University of New York. Since
1999, Dr. Bikson has authored 13 original research
articles on brain function and the effects of electric
fields. In addition Dr. Bikson has authored 5
reviews and book chapters on electric field/neuronal
interactions. Dr. Bikson is currently a member of the
CUNY Academy for the Humanities and Sciences,
the Society for Neuroscience, and the Biomedical
Engineering Society. Dr. Bikson serves on the City
College/City University of New York Medical
School Institutional Animal Care and Use
Committee.

Low Voltage Electrocution

- 15 -

M Bikson

Bibliography
Alexander L (1941) Electrical injuries of the nervous
system. J Nerv Ment Dis 94:622-632
Arkin H, Xu LX, Holmes KR. (1994) Recent
developments in modeling heat transfer in blood
perfused tissues. IEEE Transaction on Biomedical
Engineering 41(2) 97-107
Bailey B, Forget S, Gaudreault P. (2001) Prevalence
of potential risk factors in victims of electrocution.
Forensic Science International. 123: 58-52.
Biegelmeier G and Lee WR (1980) New
considerations of the threshold for ventricular
fibrillation for a.c. shocks at 50-60 Hz. Proc. Instn
Elec. Engrs. 127:103-110
Biegelmeier G (1985a) New knowledge of the
impedance of the human body in Electrical Shock
Safety Criteria, eds. Bridges JE, Ford GL, Sherman
IA, Vainberg M. Pergamon, New York, 115-132
Bridges JE, Ford GL, Sherman IA, Vainberg M
(1985) Electrical Shock Safety Criteria. Pergamon,
New York,
Bridges JE (1985) Potential distributions in the
vicinity of the hearts of primates arising from 60 Hz
limb-to-limb body currents in Electrical Shock
Safety Criteria, eds. Bridges JE, Ford GL, Sherman
IA, Vainberg M. Pergamon, New York, 61-70
Biegelmeier G (1985b) Comments in Panel
Meeting on Physiology of Electrical Stimulation in
Electrical Shock Safety Criteria, eds. Bridges JE,
Ford GL, Sherman IA, Vainberg M. Pergamon, New
York, 89-114
Bikson M, McIntyre C, Inoue M, Akiyama H, Deans
JK, Fox JE, Miyakawa H, Jefferys JGR (2004).
Effects of uniform extracellular DC electric fields on
excitability in rat hippocampal slices in vitro.
Journal of Physiology. 557: 175-190
DiMaio VJ and DiMasio D. (2001) Forensic
Pathology, Second Edition, CRC Press
Cabanes J (1985) Physiological effects of electric
currents on living organisms, more particularly
humans in Electrical Shock Safety Criteria, eds.

Low Voltage Electrocution

- 16 -

Bridges JE, Ford GL, Sherman IA, Vainberg M.

Pergamon, New York, 7-24
Camps FE, Robinson AE, Lucas BG. (1976)
Gradwohls Legal Medicine, third edition. Editors.
John Wright & Son, p360-366
Chilbert MA (1998) High-voltage and high-current
injuries in Applied Bioelectricity Reilly JR SpringerVerlag, New York, 412-453
Con Edison (2004) Second Supplemental Report of
Consolidated Edison Company of New York, Inc. In
Response to February 11, 2004 Order. Sate of New
York Public Service Commission. Case 04-M-0159
Dalziel CF (1938) Danger of electric shock.
Electrical West 80(4):30-31
Dalziel CF and Lee WR (1969) Lethal electric
currents. IEEE Spectrum Feb 1969; 44-50
Devanand DP, Dwork AJ, Hutchinson ER, Bolwig
TG, Sackeim HA. (1994) Does ECT alter brain
structure? Am J Psychiatry. 151(7):957-70. Review
DiMaio D and DiMaio V (1989) Forensic Pathology.
Elsevier, New York, p 368-369
Durand DM, Bikson M. (2001) Suppression and
control of epileptiform activity by electrical
stimulation: a review. Proceedings of the IEEE
89:1065-1082
Fatovich DM (1992) Electrocution in Western
Australia, 1976-1990. The Medical Journal of
Australia 157:762-764
Freiberger R (1934) The electrical resistance of the
human body to commercial direct and alternating
currents. Translated from German by Allen
Translation Service, Maplewood NJ, Item no. 9005.
Published by Bell Laboratories, Maplewood, NJ
Geddes LA, Baker LE (1989) Principles of applied
biomedical instrumentation, third edition. John
Wiley and Sons, New York
Gilbert TC (1939) What is a safe voltage? The
Electrical Review 119(3062): 145-146
Geddes LA, Cabler P, Moore AG, Rosborough J,
Tacker WA (1973) Threshold 60-Hz current

M Bikson

required for ventricular fibrillation in subjects of

various body weights. IEEE Trans Biomed Eng
BME-20:465-468
Gerst PH, Fleming WH, Malin JR (1966) Increased
susceptibility of the heart to ventricular fibrillation
during metabolic acidosis. Circulation Research
19(7): 63-70
Gettman KE (1985) Standards performance
characteristics in the testing of electrical fence
controllers. International Association of Electrical
Inspectors News Bulletin. July/August
Han J, Garcia P, Moe GK (1964) Adrenergic effects
on ventricular vulnerability. Circulation Research
14(6) 516-524
Hart WF (1985) A five-part resistor-capacitor
network for measurement of voltage and current
levels related to electric shock and burns in
Electrical Shock Safety Criteria, eds. Bridges JE,
Ford GL, Sherman IA, Vainberg M. Pergamon, New
York, 183-192
Hille B (2001) Ion Channels of Excitable
Membranes (3rd Edition) Sinauer Associates

Moritz AR, Henriques FC. (1947) Studies in thermal

injuries: the relative importance of time and surface
temperature in the cause of cutaneous burns.
American Journal of Pathology 23 (5): 695-720
Osypka P (1963) Quantitative investigation of
current strength, duration, and routing in ac
electrocution accidents involving human beings and
animals. Elektromedizen 8,
Peng Z, Shikui C (1995) Study of electrocution
death by low-voltage. Forensic Science
International 76:115-119.
Prieto T, Sances A, Myklebust J, Chilbert M (1985)
Analysis of cross-body impedance at household
voltage in Electrical Shock Safety Criteria, eds.
Bridges JE, Ford GL, Sherman IA, Vainberg M.
Pergamon, New York, 151-160
Rattay F (1999) The basic mechanism for the
electrical stimulation of the nervous
system. Neuroscience. 89:335-346.

Karger B, Suggeler O, Brinkman B (2002)

Electrocution autopsy study with emphasis on
electrical petechiae. Forensic Science
International 126: 210-213

Reilly JP (1998) Applied Bioelectricity. SpringerVerlag, New York

Kouwenhoven WB, Hooker DR, Langworthy OR

(1932) The current flowing through the heart under
conditions of electric shock. Am J Physiol 100: 344350
Kugelberg J (1976) Electrical Induction of
ventricular fibrillation in the human heart. Scand J
Thor Cardiovasc Surg. 10:237-240
Lee C, Meliopoulos PS (1999) A comparison of IEC
479-1 and IEEE Std 80 on grounding safety criteria.
Proc. Natl. Sci. Counc. ROC(A) 23:612-621
Leibovici D, Shemer J, Shapira SC (1995) Electrical
injuries: current concepts. Injury 26:623-627, 1995
McIntyre CC, Grill WM (1999) Excitation of central
nervous system neurons by nonuniform electric
fields. Biophys J 76:878-88.

Low Voltage Electrocution

Merrill DR, Bikson M, Jefferys JGR. (2004)

Electrical stimulation of excitable tissue: design of
efficacious and safe protocols. Invited Review
Journal of Neuroscience Methods

- 17 -

Reilly JP, Diamant AM (2003). Spatial relationships

in electrostimulation: application to electromagnetic
field standards. IEEE Transactions of Biomedical
Engineering 50(6) 783-785
Sonderdruck, Fachverlag, Schiele, and Schon, Berlin,
Translation by SLA Translation Center, TT 66-1588
& TT-11470.
Taylor AJ, McGwin G, Valent F, Rue LW (2002)
Fatal occupational electrocutions in the United
States. Injury Prevention. 8: 306-312. 2002
Taylor RJ (1985) Body impedance for transient high
voltage currents. In Electric Shock Safety Criteria.
eds. Bridges JE, Ford GL, Sherman IA, Vainberg M.
Pergamon, New York, 251-258
The Villager (2004)
http://www.thevillager.com/villager_40/familytosue
over.html

M Bikson

Webster JG (1998). Medical instrumentation

application and design, Third Edition. John Wiley &
Sons, New York
Whitaker HB (1939) Electric shock hazard as it
pertains to the electric fence. Underwriters
Laboratories, Bulletin Res. 14:3-56

Low Voltage Electrocution

- 18 -

Wright RK, Davis JH (1990) The investigation of

electrical deaths, a report of 200 fatalities. Journal of
Forensic Sciences 25:514-521.

M Bikson