Comparing Ethylene Glycol (EG) and Propylene Glycol (PG) Using Toxicity As a Basis for Risk Management Decisions. E.A. McKenna, EA, Hunt Valley, MD USA; J.S. LaKind, EA Engineering, Science, and Technology (EA), Silver Spring, MD USA; D.F. Bodishbaugh, EA, Hunt Valley, MD USA; R.P. Hubner, EA Engineering, Science, and Technology (EA), Silver Spring, MD USA; A.H. Kim, EA, Lafayette, CA USA; D.F. Ludwig, EA, Hunt Valley, MD USA; G. Wright, ARCO Chemical Company, Newtown Square, PA USA; B.S. Suedel, Dames & Moore, York, PA USA; and R.G. Tardiff, EA Engineering, Science, and Technology (EA), Silver Spring, MD USA
Ethylene glycol (EG) and propylene glycol (PG) are principal components
of aircraft deicing and anti-icing fluids and of motor vehicle
antifreeze. Consideration of the relative toxicity of these two
compounds can play a role in the selection EG versus PG formulations,
and forms the basis for managing risks at personal, regulatory,
and industrial levels. Individuals make personal purchasing decisions
based on their perception of the relative risks of such products
to humans and animals (including pets). This process may be influenced
by the common practice of many veterinarian offices to outline
the toxicity of EG-based antifreeze and promote purchase of PG-based
antifreeze to safeguard pets. At a regulatory level, both national
and international regulations for EG abound, while fewer, and
less restrictive, regulations exist for PG-based formulations.
Industries (e.g., airlines) use toxicity data to form the basis
for selecting deicing/anti-icing formulations, and a national
military (USAF) policy exists which requires the phasing out of
EG-based fluids in favor of PG-based fluids based on the former's
potential to pose a human health hazard via airfield runoff.
This review of the toxicity of EG, PG, and associated formulations
was undertaken to assess the scientific basis for supporting the
use of one over the other. The weight of the evidence presented
in this paper supports the conclusion that PG-based fluids are
preferable to EG-based fluids from a toxicological standpoint.
HUMAN FATALITIES AND ACUTE TOXICITY
Use and misuse of EG in commercial products are responsible for
a number of human deaths yearly from acute exposures. Human case
studies reporting death caused by exposure to PG have not been
found in the scientific literature.
The acute oral toxicity of EG in humans is well-documented, and
follows a defined series of stages including central nervous system
(CNS) dysfunction with severe metabolic acidosis, cardiopulmonary
failure, and acute renal failure. While the stages of acute EG
poisoning in humans have been well recorded, the stages may overlap
and death can result at any stage. The lowest reported minimum
lethal dose (MLD) of EG is 1.57 g/kg, and accidental and intentional
deaths have been attributed to ingestion of EG. Based on the
MLD, EG appears to be more toxic in humans than experimental animal
species on a weight-normalized basis. No acute inhalation exposure
studies to EG in humans have been reported, although EG caused
upper respiratory irritation, slight headaches, and lower backaches
in volunteers exposed to as much as 140 mg/m3 (37 mg/kg-day)
for 30 days. Dermal contact with EG produced mild to no skin
irritation, and minimal sensitization potential.
Acute PG toxicity, in contrast to that of EG, is not well-defined
but includes CNS effects (e.g., CNS depression) and lactic acidosis
from high doses. No deaths from oral, dermal, or inhalation exposure
to PG have been reported. Acute oral exposure to high-dose PG
has resulted in hematological effects. No acute toxicities have
been reported from the inhalation of PG vapors or aerosols. Mild
skin irritation has been reported from dermal contact with PG
in various patch tests, and the ability of PG to cause allergic
sensitization in humans remains debatable.
SHORT-TERM TOXICITY TO LABORATORY ANIMALS PREDICTIVE OF HUMAN
The acute toxicity of EG in experimental animals closely mirrors
the acute effects seen in humans, including CNS, cardiopulmonary,
and renal effects. The oral LD50 ranged from 4.0 g/kg
in rats to as high as 15.4 g/kg in mice. Rats exposed to a high
concentration of saturated EG vapor (500 mg/m3 in rats)
for 28 hours showed signs of irritation to the eyes and respiratory
tract, and mice and rats exposed to 398 Mg/M3 EG for
16 weeks (eight hours/day, five days/week) had intestinal
irritation and immune system effects. Dermal application of EG
to rabbits produced mild or no irritation.
The acute toxicity of PG includes CNS effects, hematological effects,
respiratory effects, and microscopic renal and hepatic changes
in laboratory animals. The minimum lethal dose in the rat was
19.8 g/kg and 20 g/kg in the rabbit. The LD50 ranged
from 10-20 g/kg in dogs to 33.5 g/kg in rats. No adverse effects
from inhalation of PG have been found in the scientific literature.
Dermal exposure to PG produced mild irritation.
REPRODUCTIVE, DEVELOPMENTAL AND CHRONIC TOXICITY TO LABORATORY
ANIMALS PREDICTIVE OF HUMAN RESPONSES
EG toxicity includes primarily reproductive, developmental, and
chronic renal effects. Adverse reproductive effects have been
reported following oral EG exposure of both mice and rats, with
no studies via inhalation and dermal exposures. Developmental
effects encompass developmental malformations and teratogenicity
in mice following oral exposure, with several variations and a
single malformation observed after inhalation. Dermal exposure
also resulted in developmental variations, but does not appear
to result in teratogenic effects in mice. Chronic EG toxicity
due to oral exposure results in renal effects in two species of
animals, while no studies of effects via other routes were located.
Predominant effects of non-acute PG exposure encompass primarily
hematological effects. No effects have been noted for reproductive
or developmental endpoints from exposures by the oral route.
Reproductive or developmental studies of animals exposed to PG
via other routes were not found. No chronic renal effects have
been noted, and minimal qualitative effects on the liver have
been reported. Slight hematological effects were seen in dogs
only after high-dose oral exposure and in rats after high-dose
inhalation exposure. The primary significant hematological effects
noted have been in a highly sensitive species, the cat, whose
capacity for metabolizing PG differs greatly from that of humans.
Hence, cats are not considered a representative species for modeling
PG effects in humans.
Very little chronic toxicity data exist for ethylene and propylene
glycols or glycol formulations to aquatic species. Most published
studies are best characterized as acute or subacute exposures.
Based on available data, the relative acute toxicities of EG
and PG are very similar. LC50s values for freshwater
and marine fish invertebrates and algae for EG range from 7,900
mg/L-111,000 mg/L and LC50s for PG range from 10,000
mg/L-79,700 mg/L. The acute toxicities of commercial ethylene
glycol and propylene glycol aircraft deicer/anti-icer formulations
are also similar across groups of organisms encompassing producer
(algae) and consumer species (daphnids, minnows). The range of
acute LC50 values across all taxa for propylene glycol
Type I deicers (50% diluted, deicer formulations applied hot to
aircraft to break-up ice which has already formed) is comparable
to that for ethylene glycol Type I deicers. The ranges of toxicity
reported for ethylene and propylene glycol Type II anti-icers
(non-diluted, anti-ice formulations applied cold to aircraft to
prevent ice formation prior to takeoff) are also similar, although
Type II toxicity appears to be more variable. All Type I deicers
tested are of very low acute toxicity (freshwater LC50
> 1000 Mg/L), while virtually all Type II anti-icers are more
acutely toxic (50 < freshwater LC50 < 1000 mg/L).
Based upon the limited available data, no general distinction
can be made between aquatic toxicities of ethylene and propylene
glycol formulations. There is considerable variability in the
observed toxicities of different commercial formulations, especially
those of Type II. The chemical factors which account for this
variability are unknown.