THE INFLUENCE OF SOLVENTS,
ADSORBENTS, AND CHEMICAL ANTI-DOTES ON
THE SEVERITY OF HUMAN SKIN LESIONS CAUSED BY MUSTARD GAS a
COURSE OF LESIONS
following sketch of the course of the main clinical phenomena of
lesions is given to serve as a "normal" for judging the modifications
produced by experimental
measures. Further details of the skin phenomena are contained in
Dichlorethylsulphide, as is now generally known, is a very powerful and
peculiar irritant. It
produces successively simple erythema, cutaneous edema, extensive
vesication with coagulated
contents, ulceration, and superficial sloughing. The effects, although
severe, do not extend
beyond the skin or mucous membranes.
small doses are effective. The substance is very rapidly absorbed and
fixed in the skin.
Solvents prevent the effects only if they are applied almost
immediately after contact; they are
practically useless after ten or fifteen minutes. On the other hand,
the effects develop quite
slowly. In the human skin they become perceptible only after from one
hour to two days,
according to the dosage. Histologic changes, however, start much
earlier, probably soon after the entrance of the poison.
severe burns the destructive effects grow worse for several days. In
any case the healing is
a very slow process, very much slower than with ordinary burns. Even
when healed the burned
skin remains abnormal at least for months.
striking peculiarity is the entire absence of pain or sensitiveness
during the first two
days. At this time even extensive blisters cause absolutely no
discomfort. This is not due to a
true anesthesia, for the sensitiveness to touch is unimpaired.
Apparently, the nerves are in some
way protected against the poison. A little later (on the third to fifth
day), when the necrotic
changes develop and the epidermis is more or less destroyed, the
lesions become exquisitely
sensitive to mechanical and chemical irritation. In milder cases the
edema of the skin leads to
hypersensitivity increases until the lesion has reached its
acme--between 5 to 13 days,
according to the severity. During this time the wound tends to become
covered by a sloughing
pseudomembrane. This condition persists for about a week.
toward improvement generally sets in rather abruptly between the
nineteenth day, so that within two or three days the wound
a These investigations, which
were undertaken in collaboration with the Medical Advisory
Board of the Medical Warfare Service, were carried out in the
pharmacological laboratory of the
Medical School of Western Reserve University, Cleveland, Ohio. They
were reported in the
Journal of Pharmacology and Experimental Therapeutic, Baltimore, Md..
1918, xi, No. 3, 229,
and 1919, xii, 303.
presents a healthy looking
granulating surface and the hypersensitiveness is practically gone.
This spontaneous change must be taken into consideration in judging the
success of methods of
here on improvement is steady, but slow and tedious, and likely to be
complicated by furunculosis.
63 gives the average time relations of the experimental burns. They
also hold good for
investigation upon the results of which this section is based had the
purpose of studying the
conditions that might affect the penetration and the toxicity of the
on animals were abandoned after a few trials, since their skin did not
react in the
same manner as human skin and the effects that did occur were not
TABLE 63.- Mean course of experimental lesions
urgency of the problem at the time appeared to justify direct
experimentation on human
students volunteered for what were often quite painful experiments.
experiments were, made on the author of this chapter.
poison w as applied in the form of liquid and vapor. The vapor tests
were preferred were,
whenever possible, since the results are very uniform and milder.
l-dram heomeopathic vial (about 5 cm. long and 8 mm. internal diameter)
was placed a very
small piece of cotton wool. On this was dropped 0.01 c.c. of the
poison, then another small piece of cotton, carefully wiping the mouth
of the vial. This was packed down with a glass rod and the
vial left corked for an hour to one day.
b W. P. Bowser, W. D. Cassel,
S. J. A. Foerstner, S. H Lesinger, H. H. Loucks, N. C. Wetzel.
applying, the open vial was held firmly against the skin by the thumb
and forefinger for five
minutes. The very uniform results on blank tests, and marked difference
in the presence of
efficient protectives, showed that a more complicated technique was not
normal reaction was an erythematous papule that did not go on to
0.1 c.c., was diluted with 20 c.c. of alcohol (the solution must be
within an hour, as it hydrolyzes fairly rapidly). Of this dilution
0.005 c.c. was blown from a
pipette onto the skin and spread with the point of the pipette over an
area of about one-half inch
diameter. Air was then blown through the pipette over the area, thus
evaporating the alcohol, and
leaving a thin film of the poison.
normal reaction of this dosage was a distinct vesicle which healed
slowly. This was
generally a disadvantage of the method, since it limited the number of
experiments that could be
made on a subject. It was also difficult to spread the liquid evenly
over a painted or powdered
first stronger doses were used (0.01 c.c. of 1 percent), but these
sometimes gave rise to
this, 0.1 c. c. of dichlorethylsulphide was spread on filter paper so
as make an even stain,
about 1 by 5 cm. This was then cut into 10 squares of 5 mm. each, so
that each square
represented about 0.001 c. c. of dichlorethylsulphide. These were laid
on the skin and kept in
place by a bandage for one or two hours.
reaction was altogether too severe, and this method was tried only in
connection with fairly
lesions were inspected at suitable intervals and frequently
EFFECTS OF SOLVENTS
first line of experimentation concerned the effects of water and oil.
readily soluble in oil, and very little soluble in water. It therefore
seemed probable that these
solvents would affect its penetration and the irritation in the same
way as they do phenol or true
mustard oil. For these it was found that the irritant distributes
itself between the solvent and the
skin, according to its partition coefficient; the greater its affinity
for the solvent the slower will
be its penetration into the tissues, and the smaller the irritation,
and vice versa.
results and experiments showed that the same principles apply to
Waiter increases the irritation (Fig. 223) and oils render it less
irritant (Fig. 224).
however, the absorption is still very rapid, even in the presence of
oils, and the
ultimate injury is not very much reduced. The importance of the delayed
absorption is further
minimized by the fact that the irritation is proportional to the
absolute quantity rather than to the
concentration of the poison. This difference from most other irritants
is explainable on the
assumption that the toxic
effects are not due to the dichlorethylsulphide itself, but to its
intracellular decomposition, resulting in the liberation of
hydrochloric acid within the cells. The
degree of this intracellular acidosis would, of course, depend upon the
absolute quantity of the
dichlorethylsulphide that had penetrated into the cell. Since the
poison probably does not leave
the cells after its absorption, the rate of absorption would have
little effect on the absolute
these restrictions, however, the influence of the solvents is quite
certain limits. The protection by oils is especially efficient in
prolonging the time during which
removal treatment remains effective. (Fig. 225.)
efficiency also increases with the thickness of the oily layer on the
skin; and this depends
largely on the stiffness of the oil or ointment. The efficiency is
therefore increased by the
addition of powders or "fillers."
FIG. 223.- Detrimental
effects of water (vapor tests). Dichlorethylsulphide vapor was applied
the skin; 62. bare skin; 63, skin coated with water; 64, skin coated
with sodium bicarbonate
water paste; CZ, skin coated with soap; 66, skin coated with kaolin
water paste. The photograph
was taken 7 days after the application. Note the lesion is greatest on
the moistened skin (63), as
compared with the bare skin (62). Kaolin restrains this somewhat (66).
Sodium bicarbonate paste
(64) is useless. soap paste (65) is fairly efficient. (The figures in
parentheses refer to
effect appears to be purely mechanical, for no material differences
exist between the
various substances that were tried.
are, however, differences between the oils themselves that are not
explainable on a
physical and mechanical basis, and must therefore presumably be
chemical. Linseed oil, raw as
well as boiled, and cod liver oil furnish considerably more protection
than do other oils. (Fig.
226.) This is perhaps due to unsaturated fatty acids. The efficiency,
however, is also limited. (Fig. 227.)
certain conditions, the fats apparently may increase the toxicity:
namely, when they
facilitate the contact with the poison. This occurs, for instance, when
cloth saturated with the oil
is laid loosely on the skin. (See fig. 224.)
following gives a more detailed outline of the experiments and results:
effects of dichlorethylsulphide are more severe in the presence of
moisture on the skin. This
is shown very convincingly by the vapor test, in experiment No. 63.
(See fig. 223.)
action of water is reflected in the attempt to use protective
substances as watery solutions or
pastes. These are uniformly less effective than the dry substances, and
in nearly all cases they are
also less effective than the oily solutions or pastes.
of their water content, the following rendered the skin
hypersusceptible, so that the
lesions were more severe than on the bare skin: 50 percent glycerin
(experiment No. 15); sodium
bicarbonate-water paste (experiment No. 64); kaolin-water paste
(experiment No. 66); fuller's
earth-water paste (experiment No. 68).
deleterious effects of moistening the skin do not contradict the
beneficial effects of washing
and scrubbing the skin, after exposure, with soapy solution.
FIG. 224.- Protective
action of petrolatum when dichlorethylsulphide is applied as "splash,"
when applied through fabric. In experiments No. 10 and No. 11, the
dichlorethylsulphide was applied directly to the skin; No. II was
previously vaselined. The
protective value of the petrolatum in No. 11 is apparent. In No. 12 and
No. 13 the alcoholic
dichlorethylsulphide was dropped on a small square of cloth and this
was applied to the skin; No.
13 was frst oiled with petrolatum. Note that the two lesions are
practically alike. The photograph
was taken 19 days after the application
other hand, it is evident that the skin should be protected against
exposure by keeping it covered with absorbent dusting powders; or
better, by keeping it oiled, if
that is practical.
SIMPLE OILS AND FATS
oils restrain the effects of dichlorethylsulphide. They differ
relative protective efficiency can be seen when the vapor or alcoholic
solutions are applied
to the oiled skin; but they are especially striking if equal doses of a
3 percent solution of
dichlorethylsulphide in various oils is applied to the skin. The
protective efficiency in the
different series is in the following order, the most effective
protection being at the top, the least
effective at the bottom of the list.
VAPOR TESTS C
Boiled linseed oil
(53). Castor oil (90).
Liquid petrolatum (47).
numbers in parentheses refer to experiments.
DICHLORETHYLSULPHIDE APPLIED TO OILED SKIN d
Raw linseed oil (117).
Liquid petrolatum (2, 6,
7). Solid paraffin (122).
Olive oil (3).
DISSOLVED IN OILS (3 PERCENT)
Raw and boiled linseed oil
(111 and 112) Olive oil (113).
Castor oil (114)
(See Fig. 226.)
Liquid petrolatum (110).
FIG. 225.- Value of
protective oiling. In No. 8 the dichlorethylsuliphide was applied to
the hare skin, in No. 9 to oiled skin. Both were washed with oil after
15 minutes. Observe the much
greater effect on the unprotected skin in No. 8. The photograph was
taken 24 days after the
these data the general order of efficiency is: Linseed oil, raw or
boiled, cod-liver oil, solid
paraffin, petrolatum, liquid petrolatum, olive oil, castor oil,
The oiling of the skin is
decidedly protective against slight exposure (experiments 47, 53, 90)
to the vapor and fairly
effective against small doses of the liquid (2, 3, 4, 11). Its
usefulness, however, is not unlimited.
Even the most effective oils do not prevent blisters if the strong
dichlorethylsulphide is left in
prolonged contact (117 to 122). (See fig. 227.)
the substances that were tried as chemical antidotes really acted
merely as fillers. The
tabulations are again arranged in the order of efficiency, those giving
the most complete
protection are at the top; the least protection at the bottom. Some of
the plain oil are given in
brackets for comparison.
VAPOR TEST (SOLLMANN)
Liquid petrolatum stiffened with charcoal (50), kaolin (49), or
d The numbers in parentheses
refer to experiments.
VAPOR TEST (CASSEL)
Zinc oleate (97)
and solid petrolatum pastes made with:
Silver abietate (resinate) (101).
(95).Boric acid (99).
(96).Solid petrolatum (100).
Manganese dioxide (98).
Zinc in. liquid petrolatum
Hexamethylamin in liquid
petrolatum (20), (petrolatum, 11).
the dichlorethylsulphide is concentrated, even the filled ointments
have only slight
values. This is shown by the series 116 to 122, which also includes
solid paraffin. (Fig. 227.) In
such cases only the chlorine preparations are promising.
FIG. 226.- Comparison of
oils. Observe the degree of effect. Each area received 0.01 c. c. of 3
percent solution of dichlorethylsuiphide in the oil, spread over a
surface of about one-half inch
diameter. The photograph was taken 2 days alter application. The
solvent oils were as follows:
110, liquid petrolatum; ill, raw linseed oil; 112, boiled linseed oil;
113, oilve oil; 114, castor oil;
115, codliver oil. (The figures refer to experiments.)
PROTECTIVE VARNISHES e
may be supposed to act like the fats. They would remain longer on the
skin, but they can
only be applied in a very thin coat, and this is likely to furnish only
an incomplete protection. A
thicker coating might soon become harmful.
actual tests by the vapor method, although not complete, were not
paint (83) actually increased the irritation; asphalt-ether varnish
(80) was no better
than bare skin. Collodion (85) furnished a very slight protection.
following protected, but it was not determined whether the protection
was greater than that
of simple oiling: Rosin-ether varnish; also with zinc stearate and with
fuller's earth; shellac
e The numbers in parentheses
refer to experiments
were found highly effective, just as they are in the masks. A layer of
1 mm. thick protects
completely against the vapor test (Fig. 228) and quite effectively
against the alcoholic solution.
The results were as follows:
Vapor tests on dry powders (Sollmann). Bare skin erythematous papule. No lesion followed on
cocoanut charcoal (29). Slight and inconstant erythema (less than
linseed oil 53): Kaolin (27,
73), fuller's earth (28, 74). Slight papular erythema: Zinc stearate
(30) (about like liquid
Vapor tests on dry powders (Loucks)
.-Bare skin erythematous papule. No lesions followed on:
Manganese dioxide (103), talcum (108), zinc oxide (105). Slight
erythema: Silver abietate (107);
reduced iron (104), litharge (106). Moderate erythema: Zinc stearate
Liquid test with alcoholic
dichlorethylsulphide.- Fuller's earth protected most; talcum,
intermediate protection; calcium carbonate, least protection.
FIG. 227- Protective value
of dichloramine-T. Pieces of filter paper about 5 mm. square, and
containing about 0.001 c.c. of 95 percent dichlorethylsulphide were
applied to the center of a
square of skin, covered with the protectives. The coatings were as
follows: 117, raw linseed oil;
118, linseed oil and kaolin, I to 3; 119, same with 3 percent of soft
soap; 120, petrolatum kaolin,
I to 1; 121, dichloramine-T, 10 percent in chlorcosane; 122, solid
paraffin. The photographs were
made I day after the application. Note that the reaction is less in 121
than in the others; 26 and 30 are recrudescences of lesions 25 days
old. (The figures refer to experiments.)
results show that the general efficiency of the powders is as follows,
the most effective
being at the beginning, the least effective at the end of the list:
Cocoanut charcoal; fuller's earth,
kaolin; talcum, manganese dioxide, zinc oxid; silver abietate, reduced
iron, litharge; zinc
stearate; calcium carbonate.
the usefulness of the adsorbent powders is limited by the difficulty of
on the skin in sufficient thickness.
above tabulation of relative efficiency indicates plainly that the
following act only
mechanically, and not chemically, and that this efficiency is actually
lower than that of the
cheaper charcoal, kaolin, or fuller's earth, namely, metallic soaps
(zinc stearate), metallic
resinate (silver abietate), metallic oxids (zinc oxid and litharge),
and free metals (reduced iron).
ADSORBENT WATER PASTE
attempt was made to secure a better adhesion of the adsorbents to the
skin by using them in
the form of pastes. These pastes are not nearly as efficient as the dry
powders. This is
explainable partly by the deleterious effects of water itself, and
partly by the watery film,
preventing ready access of the nearly insoluble dichlorethylsulphide to
following experiments were made, the results again being presented in
order, the most
efficient abcove, the least effective below.
Vapor tests (Sollmann).- Bare skin
gives erythematous papule.
Inconstant slight erythema: Cocoa-charcoal paste (67) (about like dry
erythema, rather less than on bare skin: Powdered zinc paste (71).
on bare skin: Fuller's earth paste (68).
more severe than on bare skin: Kaolin paste (66).
vesication: Sodium bicarbonate paste (64).
alone produces more severe blister and scabbing (63).
Vapor tests (Cassel).- Collargol,
10 percent (93) did not furnish protection, giving the same
results as on the bare skin.
Liquid alcoholic dichlorethylsulphide.-
Alkresta (purified fuller's earth) in 50 percent glycerin
(16) furnishes a fair protection, while 50 percent glycerin alone
increases vesication (15).
FIG. 228.- Efficiency of
dry powders. Dichlorethylsulphide vapor was applied. Note the normal
lesion in the bare skin (26); the slight protection afforded by zinc
stearate (30); and the complete
protection from kaolin, fuller's earth, and charcoal (indicated by the
absence of lesions in a
straight line drawn from 26 to 30). The photograph was taken 3 days
after the application. (The
numbers in parenthesis refer to experiments.)
soap" furnishes considerable protection, chiefly, it is presumed, by
acting as a
solvent. It acted as well as linseed oil or zinc oleate in the vapor
test (65). It may also be
incorporated with the ointments; for instance, 5 percent with
petrolatum (91), and 3 percent in
the linseed kaolin mixture (119). The addition of the soap modifies
ointments so that they can be
applied more smoothly and also facilitates their removal by washing.
however, are not very great, and it is conceivable that the alkalinity
of the soaps might injure
Certain reports by other workers indicated that metallic soaps and
would be especially effective. This was not confirmed. Those tried had
only at limited success,
and this was attributable mainly to their mechanical action as
absorbents and fillers.
Lead plaster (54).- Protected well
in the vapor test, but this was explained by the dense
consistence of the film.
Zinc oleate (the
old U. S. P. preparation).- Protected partially in the vapor test,
being about equal
to linseed oil (53) and very little better than kaolin- petrolatum
ointment (50). On another subject
(97) it protected completely, but so did ointment of zinc oxide or
Zinc stearate in the dry form (30,
107).- Protects very little, even against vapor; much less than
dry kaolin (28), talcum (108), or zinc oxide.
petrolatum ointment.- Furnished protection against vapor (95),
but so did similar
ointments of zinc oxide (96) and boric acid (99).
Silver abietate (resinate), dry.-
This furnished only a relatively slight protection against vapor
(107), being inferior to talcum (108).
Silver abietate-petrolatum ointment.-
This protects against the vapor (101), but so did a similar
boric acid ointment (99).
COLLOIDAL METALS AND METALLIC OXIDES
were tried in the hope that they might act as catalysts, accelerating
the hydrolysis of the
dichlorethylsulphide. The results show that this does not occur under
the conditions of their use
on the skin. They furnish some protection, but no more than any
indifferent powders. The
following were tried:
In dry powderform.- Powdered zinc
(79): Protection against vapor much less than dry kaolin
(73). Manganese dioxide (103) and zinc oxide (105): Protect against
vapor, but so did talcum
As water pastes.- Zinc dust paste
(71): Protects very little better than fuller's earth paste (68),
both being about the same as on the bare skin (62). Collargol, 10
percent (93): Does not protect
ointments.-Zinc dust ointment: Protects somewhat better than
probably not better than any indifferent powder, such as
hexamethylanamin (20). Manganese
dioxide ointment (98), zinc oxide oint- ment (96), and collargol
ointment (94): Protect against
vapor, but so did boric ointment (99).
caustic action of dichlorethylsulphide is destroyed by further
chlorination. Chlorinated lime
and the chloramines are effective under working conditions. Their
efficiency is limited; and their
practical application is further confined by their irritant action on
the skin, the instability of
some of the preparations, and the cost of the chloramines.
efficiency is illustrated by Figure 227 and by the following
In the vapor tests.- Complete
protection was secured by: Dichloramine-T, dry (76), or in water
paste (69), or as 10 percent dusting powder (78) and 10 percent in
chlorcosane. Calx chlorinata,
in the same forms (75, 70, 77). The efficiency of dichloramine-T and
calx chlorinate was equal
to that of dry charcoal, and superior to all others. Chloramine-T
paste, Squibb (52), gave almost
Against concentrated liquid G34.-Dichloramine-T,
10 percent in chlorcosane gave only partial
protection (121), but was somewhat more efficient than linseed kaolin
pastes (118, 119). This is
shown in the Figure 227.