Brit. J. Pharmacol. (1964),23,455-475. COBALT COMPOUNDS AS ANTIDOTES FOR HYDROCYANIC ACID BY C. LOVATT EVANS From the Chemical Defence Experimental Establishment, Porton Down, Salisbury, Wilts. (Received February 18, 1964) The antidotal potency of a cobalt salt (acetate), of dicobalt edetate, of hydroxo- cobalamin and of cobinamide against hydrocyanic acid was examined mainly on mice and rabbits. All the compounds were active antidotes for up to twice the LD50; under some conditions forlarger doses. The most successful was cobalt acetate for rabbits (5xLD50), which was effective at a molar cyanide/cobalt (CN/Co) ratio of 5, but had as a side-effect intense purgation. Hydroxocobalamin was irregular in action, buton the whole was mosteffective for mice (4.5XLDS0 at a molarratio of 1), and had no apparent side effects. Dicobalt edetate, at molar ratios of up to 2, was more effective for rabbits (3xLD50) than for mice (2xLD50), but had fewer side effects than cobalt acetate. The effect of thiosulphate was to augment the efficacy of dicobalt edetate and, in mice, that of hydroxocobalamin; but, apparently, in rabbits, to reduce that of hydroxocobalamin. Cobinamide, at a molar ratio of 1, was slightly more effective than hydroxocobalamin on rabbits and also less irregular in its action. Cobalt acetate by mouth was effective against orally administered hydrocyanic acid. The oxygen uptake of the body, reduced by cyanide, is rapidly reinstated when one of the cobalt antidotes has been successfully administered. Hydrocyanic acid acts as a poison because it combines with, and immobilizes the function of, the iron atom in cytochrome oxidase (derived from cytochrome a3) (Keilin & Hartree, 1939), so blocking the electron transfer through the cytochrome system, and checking the final oxidations involving oxygen uptake of the tri- carboxylic acid cycle. It also inhibits many other enzymes, for example, decarb- oxylases (Blaschko, 1942); these last effects have no recognizable relation to the toxic action. Attempts to find an antidote for cyanide have usually been in the direction of breaking down the cyanide-cytochrome a3 complex, and have looked for substances with a stronger affinity for cyanide ions than has the oxidase. One such method is to convert some of the blood haemoglobin into methaemoglobin, for example, by injection of a nitrite (Paulet, 1959); the methaemoglobin combines with the cyanide ion to form the very stable cyanmethaemoglobin. This has the drawback that it involves the loss of oxygen-carrying power ofthe blood, so that, for example, to antidote twice the LD50 of cyanide would be equivalent to the loss of about 10% (or in an adult man 450 ml.) of the blood. Since cobalt salts form stable complexes with cyanide, the cobaltocyanides, M4Co(CN)6, and the cobalticyanides, M3Co(CN)6, they have been shown long ago 456 C. LOVATT EVANS (Antal, 1894; Lang, 1895; Meurice, 1900; Evans & Watt, 1942, unpublished; and Mercker & Bastian, 1959a, b) to be efficient antidotes. In vitro each mole of cobalt can fix up to 6 moles of cyanide, and if that holds for the conditions in vivo, relatively small amounts of cobalt salts would be needed for the neutralization of several lethal doses of cyanide. Another substance which combines firmly with cyanide is hydroxocobalamin, or vitamin B12a (Conn, Norman & Wartman, 1951), cyanocobalamin or vitamin B,2 being formed thereby, and this is stable in the dark, but is decomposed again, yielding the hydroxo-compound in light. Braekkan, Njaa & Utne (1957) concluded that the liver stores the hydroxocobalamin in preference to the cyanocobalamin. In the body, small amounts of hydrocyanic acid present in the blood convert the hydroxocobalamin into vitamin B12 and Wokes & Picard (1955) believe that this reaction plays an important part in endogenous cyanide metabolism. Vitamin B12 functions in the body in the form of a coenzyme (aS ;6-dimethylbenzimidazole- cobamide), in which an extra sugar and an extra adenine form an attached nucleoside (Lenhert, 1962). The antidotal action of hydroxocobalamin against cyanide has been studied by Mushett, Kelly, Boxer & Rickards (1952), Mercker & Bastian (1959b), Delga, Mizoule, Veverka & Bon (1961), Delga, Mizoule & Veverka (1961) and Paulet, Bernard & Olivier (1963), and the compound was found to be effective. Various other cobalt complexes have also been tried, notably dicobalt edetate (the chelate of two cobalt atoms with ethylenediamine tetracetic acid), of cobalt with histidine, and also the gluconate and glutamate of cobalt (Paulet, 1957, 1958, 1960; Mercker & Bastian, 1959a, b; Bartelheimer, 1962b; Tauberger & Klimmer, 1963). This paper describes a study and comparison of the antidotal actions of cobalt salts, hydroxocobalamin, the hydrolysis product cobinamide, and dicobalt edetate, with a view to providing information that may be of use in cases of poisoning with hydrocyanic acid or cyanides in industry, which now uses these substances in very large quantities. METHODS The LD50 ofhydrocyanic acid, by various routes of administration, was determined for mice and rabbits. Alkali cyanides were not used as such, as the solutions are so strongly hydro- lysed as to be very alkaline (for example 0.02 M solium cyanide has a pH of 10.7 and is 98% ionized), and to cause pain on injection, and subsequent tissue damage. Instead, a solution of sodium cyanide was neutralized with acetic acid to about pH 7, the content of hydrocyanic acid was estimated by silver titration, and the subsequent dilution to the required concentration was made with 0.9% saline. At pH 7 the hydrocyanic acid is about 1% ionized. In some experiments pure hydrocyanic acid was used, but this had no advantage and was more difficult to handle. Owing to the volatility of hydrocyanic acid, the solutions were kept in well- stoppered vessels and, when any doubt existed as to the final concentration, were again titrated at the end of the day's experiments. It was usually found that after 4 hr the concentration had fallen, often by 2 to 5%, and sometimes in hotweather by as much as 10%. For the solution of cobalt salts the acetate (recrystallized) was chosen, as its solution in water is less acid than that of the chloride or nitrate. The pH of 0.1 mM solutions were for acetate 7.42; for chloride 5.78, and for nitrate 6.11. COBALT COMPOUNDS AS CYANIDE ANTIDOTES 457 Intravenous injections were made into the tail vein in mice and an ear vein in rabbits; intra- muscular injections were into the thigh muscles. Doses are expressed usually in terms of umoles/kg of body weight, and for hydrocyanic acid sometimes also in terms of LD50. Owing to the rapidity of action of hydrocyanic acid when large doses are given, and to the difficulty of giving intravenous injections when the animal is convulsing, it was more usual to give the antidote first, intravenously, and then to follow this in a matter of seconds by the hydrocyanic acid, usually by intraperitoneal or intramuscular injection, but in some experi- ments that order was reversed; the antidotal effects were at least as great, and often greater, when that was done, provided the antidote was given soon enough. In order to indicate the molar relationship between the hydrocyanic acid and the antidote in each experiment, the ratio between them was usually expressed (Imoles of hydrocyanic acid per kg divided by eLmoles of cobalt compound per kg) and is called the cyanide/cobalt ratio. RESULTS Toxicity of hydrocyanic acid By intravenous injection. The LD50 for mice was found to be about 40 Etmoles/ kg, and for rabbits 30 ttmoles/kg. Mice given two- to three-times the LD50 convulse in a few seconds and die in 1 to 2 nun. By intraperitoneal injection. As expected, the results showed a rather large spread, the mean LD50 for rabbits being 58 ,umoles/kg, and for mice 111 Etmoles/ kg, so that by this route mice were only about half as sensitive as rabbits. This is probably due to a more rapid detoxification by the mouse liver. By other routes. It would be expected that toxicity would be maximal when the cyanide is given quickly into a vein which drains into the inferior vena cava, so that it is quickly delivered to the central nervous system, less toxic when it enters more slowly, as by intramuscular injection, and least toxic when it enters the circulation slowly, as when given intraperitoneally, subcutaneously or orally. And, in fact, the descending order of toxicity proved to be intravenous-intramuscular- intraperitoneal-subcutaneous-oral, as shown in Table 1, which also includes some results by other authors. The results of inhalation are more vague, but are probably comparable with those by intravenous injection. A method of administration often used (Paulet, 1955a, b, 1960; Delga et al., 1961a, b), but not employed in the present investigation, is to anaesthetize the animal and to inject the cyanide intravenously at a constant slow rate of about 0.1 mg/kg/min. The results so obtained do not seem to differ greatly from those of the usual procedure for intravenous injection. When the first period of apnoea occurs the dose received is lethal, that is, if the infusion is stopped then, the animal will nevertheless die. These results refer to dogs. Mice do not restart respiration as a rule, after apnoea. The toxicity of cobalt salts The cobaltous ion is known to be toxic to some micro-organisms, to depress metabolism in tissue slices (Burk, 1946) and, under certain conditions, metallic cobalt can be carcinogenic (Heath, 1960), though this last action has not been demonstrated for the cobalt ion. 458 C. LOVATT EVANS TABLE I TOXICITY OF HYDROCYANIC ACID I.v., intravenous ; i.m. intramuscular; i.p., intraperitoneal; s.c. subcutaneous. *At 0-1 mg/kg/min 95/o Fiducial Routeof LD50 limits LD50 Species administration (,umoles/kg) (jumoles/kg) (mg/kg) Reference Rabbit I.v. c.30 0-82 Thispaper L~v.* c.27 07 Delga et al. (1961a, b) I.m. 407 303-52 1 10 Thispaper I.m. 55 1.5 Meurice (1900) I.p. 58-0 490-70 1 57 Thispaper S.c. c.930 2-5 This paper, Lang (1895) Mouse I.v. 400 11 Thispaper I.v. 67-0 1P75 Paulet (1955a, b, 1960) I.m. 100-0 2-7 This paper I.p. 111.0 94-124 2-99 Thispaper Oral 155-0 4-17 Thispaper Oral 140 3-8 Delgaet al. (1961a, b) Rat S.c. 137 3.7 Bartelheimer (1962 b) In acute experiments, cobalt salts given intravenously cause a fall, followed by a rise anda later fall, of arterial pressure, an increase in breathing, cramps, vomiting and acute diarrhoea with intestinal inflammation (Stuart, 1884; Le Goff, 1930; Bartelheimer, 1962a; Tauberger & Klimmer, 1963). These effects have been analysed by Bartelheimer (1962a), and by Tauberger & Klimmer (1963), and shown to be partly due to central action and partly to peripheral effects. Death is due to respiratory failure when the administration is rapid, to cardiac failure when more slowly given (Bartelheimer, 1962a). (a) (b) (c) 0 C * S NaCN CoAc NaCN CoAc Fig. 1. RabbitileumpreparationinoxygenatedTyrodesolution,inorganbathof25ml. (a)addition of1mg(20-5,umoles)ofsodiumcyanide(NaCN) ; (b)afterwashingadditionof1mg(4,umoles) ofcobalt acetate (Co Ac); (c),afterwasbingadditionof1 mgofsodiumcyanidethen 1 mg ofcobaltacetate. Cyanide/cobalt molarratio = 5-2. Time scale, I min. COBALT COMPOUNDS AS CYANIDE ANTIDOTES 459 In more chronic administrations the principal effects are polycythaemia, porphyrinuria, increase in the size of the adrenals, and goitre (Gairdner, Marks & Roscoe, 1954; Saikkonen, 1959). Cobalt salts have been given orally to human subjects in doses up to 150 mg of cobaltous chloride daily without clearly pro- nounced harmful effects (Davis & Fields, 1955). Orally administered cobalt is very slightly absorbed, and mainly excreted in the faeces; injected intravenously it is mainly, and rapidly, excreted in the urine (Taylor, 1962); according to experiments with radioactive 60Co (Cuthbertson, Free & Thornton, 1950) what remains in the body is to be found mainly in liver, kidneys, pancreas and spleen. In the present experiments, no histochemical evidence could be found that cobalt was present in any tissues or organs 12 hr after intravenous injection. Post mortem examination showed intense congestion of the mucosa of the intestinal tract. On the rabbitisolatedileum preparation the effect was relaxation, withdiminution of the rhythmic contractions (Fig. 1); on the rectum in situ intravenous injection caused powerful rhythmic contractions (Fig. 2), and these probably account for the colic and diarrhoea which often follow administration of cobalt salts. Respiration Rectal Contractions 140 -0E~~ Co Ac Fig. 2. Cat, anaesthetized with pentobarbitone sodium. Uppermost tracing, respiration middle tracing,contractions ofrectum; lowesttracing,arterialbloodpressure.Atthemaik20mg/kg (pmoles/kg) ofcobalt acetate (Co Ac) was injected intravenously. Timescale, 1 min. The toxicity of cobalt compounds is related to the ease with which they yield cobalt ions, and hence is greatest in salts of cobalt. Cobalt acetate yields ions readily; Siddhanta & Banerjee (1958) give the dissociation constant for Co.O.CO.CH3+ 'Co+++O.CO.CH3- " as 30x 10-3, or pK=1.8, and the compound would be expected to be somewhat toxic. The chelate compounds of cobalt, which will be referred to later, do not yield ions so readily, if at all. The lethal doses of various cobalt compounds are given in Table 2. 460 C. LOVATT EVANS TABLE 2 LETHAL DOSES OF COBALT COMPOUNDS EDTA=editicacid ; i.v., intravenous ; i.p. intraperitoneal LD50 Species Compound Route (. moles/kg) (mg/kg) Reference Mouse Coacetate I.v. 125 31 Thispaper Rabbit Coacetate I.v. 100 25 Thispaper Rat CoCI2 I.v. 83 20 Tauberger&Klimmer(1963) Cat Co Cl2 I.v. 120 27 Bartelheimer(1962a,b) Mouse CoNaEDTA I.p. 6,700 1,948 Eybletal. (1959) Mouse Co2EDTA I.v. 122 50 Paulet(1955a, b, 1961) Mouse Co2EDTA I.v. 175 71 Thispaper Rat Co2EDTA I.v. 106 43 Tauberger&Klimmer(1963) Rat Co,EDTA I.p. 100 41 Bartelheimer (162a, b) Rat Co-histidine I.v. 287 104 Tauberger&KIinumer(1963) Rat Co-histidine I.p. 370 134 Bartelheimer (IS62a, b) Cat Co-histidine I.v. 136 50 Tauberger&Klimmer(1963) Cobalt salts as antidotes Exactly what occurs when cobalt and cyanide ions are introduced into the body is uncertain; initially the cobaltocyanide ion (Co(CN)6---) would be formed, but the ultimate product is probably cobalticyanide (Co(CN)6--) which has been shown by Bartelheimer, Friedberg & Lendle (1962) to have a log stability constant (log Ks) of 19, so it is very stable, while being at the same time only slightly toxic (LD50=1,000 mg/kg). Cobalt ions can reach the brain, that is they can penetrate the blood-brain barrier (Smith, 1962; Bartelheimer, 1962a), have no precipitating action on proteins, and, when introduced into the blood stream, are eliminated by the kidney, which is not damaged. The antagonism between cobalt and cyanide ions can be illustrated by experi- ments with isolated tissues in vitro. For example, as shown in Fig. 1, on the isolated small intestine the action of 20.5 jumoles of cyanide (giving a concentration of 0.82 jumoles/ml.) was eliminated by 4 pmoles of cobalt acetate, the cyanide/ cobalt ratio being 5.2. Some experiments were made to find whether cobalt ions were able in vitro to reverse the inhibition of the cytochrome oxidase system produced by cyanide. using p-phenylenediamine as the oxidation substrate. Only about 10% of the initial oxidase activity was restored, however, probably on account of the facts that phosphate buffers were used and the insolubility of cobalt phosphate, and these experiments were not pursued further. Since the lethal dose of cobalt salts is around 100 umoles/kg, the amount given as an antidote has mostly been lower than that; but that dose should be able to antidote 600 umoles of hydrocyanic acid, if the theoretical ratio holds good, and the mutual antidotal action should in theory reduce any risk of cobalt poisoning. This would mean that, for a safe dose of a cobalt salt, some 10xLD50 of hydro- cyanic acid intraperitoneally for rabbits, and 5xLD50 intraperitoneally for mice, should be the maximal amounts that could be antidoted. In experiments done in 1942 (Evans & Watt, unpublished) the maximal dose antidoted was 10x LD50, in a goat. The present series, using mice and rabbits, gave results which are satis- COBALT COMPOUNDS AS CYANIDE ANTIDOTES 461 TABLE 3 THE EFFECTS ON RABBITS OF COBALT ACETATE AND HYDROCYANIC ACID Thecobalt acetate was given intravenously; thehydrocyanic acid intravenously (i.v.) or intraperi- toneally (i.p.). The hydrocyanic acid was given first, except for asterisked results, when the cobalt acetate was given first. Figures in brackets after the number ofrabbits died give the numbers of rabbits tested. *Indicates thefirsttobeadministered Hydrocyanic acid Cobalt Cyanide/ Number died Dose Multiple acetate cobalt (imoles!kg) Route ofLD50 (jumoles/kg) Ratio Total % 178 I.v. 6-0 31* 5 8 (1) 0 170 I.v. 5-7 84 2-0 0(1) 0 0 167 I.v. 5 6 100 1P7 0(2) 0 148 I.v. '50 30 50 (1) 0 0 133 I.v. 4-4 100 1P3 0(1) 0 119 I.p. 20 10* 12-0 1 (1) 100 116 I.v. 3-8 14 8-3 4(4) 100 116 I.v. 3-8 23 5.1 0(2) 0 85 I.v. 2-8 84 1P0 (5) 74 I.p. 1P3 42 1P8 00(1) 00 60 I.V. 2-0 9 6-6 (1) 0 0 60 I.V. 2-0 19 3-2 1 (1) 100 54 I.V. 1-8 17* 3-2 0(1) 0 factory for rabbits, but for mice fall short of the theoretical predictions mentioned above, as is shown in Tables 3 and 4. When the molecular ratios of cyanide and cobalt are calculated, it is seen that, in rabbits, the ratios are spread over the range 1 to 12. If it is assumed that each mole of cobalt can deal with six of cyanide, then in all cases except two, where the cyanide/cobalt ratio exceeded 6, there should be no hydrocyanic acid left free. In the two instances where an excess would be left it would be only a fraction of an LD50, and this should have been nonlethal, but in one instance (cyanide/cobalt ratio of 12) was not, which indicates that the case is not so simple as it first seemed. The results in Tables 3 and 4 show that the cobalt salt was more effective as an antidote for rabbits than for mice. For rabbits, provided the molar cyanide/ cobalt ratio does not exceed 6, it seems that at least 6xLD50 can be neutralized, while at a ratio of 12 even twice the LD50 is not antidoted. Ir should be noted, TABLE 4 EFFECTS ON MICE OF COBALT ACETATE FOLLOWED BY HYDROCYANIC ACID Cobaltacetatewasgivenintravenously,and hydrocyanic acidintraperitoneally. Figures in brackets afterthenumbersofmicediedgivethetotalnumbersofmicetested Hydrocyanic acid Cobalt Cyanide/ Numberdied Dose Multiple acetate cobalt (G.moles/kg) ofLD50 (pmoles/kg) Ratio Total % 450to 4-1 to 75 to 2-0to 20(20) 100 900 8-2 225 6-0 300 2-7 62 4-8 10(11) 91 250 2-25 50 5*0 (5) 100 225 2-0 75 30 58 (9) 89 225 2-0 46 4.9 20(32) 63 222 2-0 25 9*0 (5) 100 220000 11--88 3530 46--00 053 ((49)) 750 150 135 3'0 0(8) 50 0 462 C. LOVATT EVANS however, that, inthe rabbit series, hydrocyanic acid was usually given intravenously, whether before or after the cobalt. With mice the picture is much less satisfactory; they are less sensitive to hydro- cyanic acid, and about as sensitive to cobalt, as are rabbits, and the results are less regular. It appears from the results in Table 4 that, under the most favourable conditions, only upwards of twice the LD50 can be antidoted, and then only when the cyanide/cobalt ratio is below 5. The hydrocyanic acid was given intraperi- toneally and the cobalt solution intravenously in this species, which might have affected the results. Dicobalt edetate as antidote Experiments were carried out with pure dicobalt edetate., and also with the solution sold under the name of Kelocyanor (Laboratories Laroche Navarron, 63 Rue Chaptal, Levallois, Paris, Seine, France) which is provided in 20 ml. ampoules, each containing a 1.5% solution in 20% glucose. The solution has a pH of 4.42. The dicobalt edetate and the cyanide ion have been shown by Paulet (1960) to be mutual antidotes, the antagonizing amounts found by him being 0.1 mg of cyanide:0.8 mg of dicobalt edetate (or 3.9 ptmoles: 1.96 jumoles, or 2:1, andnot 6: 1 as might have been expected if the second cobalt atom ofthe compound were fully ionized). The toxicity of the compound, given in either form, was found to be about 175 ptmoles/kg for mice (Paulet, 1960, found about 122 tmoles/kg) so, in spite of a probable mutual antidotal action, the amount given was usually kept below that dose. Itwas tried with and without simultaneously intravenously administered thiosulphate, and both before and after the cyanide. Results are given in Table 5. Thiosulphate alone had antidotal action which varied; for mice the value was around 1.5 to 2xLD50 at the best, but was irregular, and in rabbits it was also around twice the LD50. The results are rather irregular, but show that, for mice, some 2 to 3xLD50 can be antidoted when the cobalt compound is given in the form of Kelocyanor, that is, with glucose, and when thiosulphate is also given. The pure dicobalt edetate was less successful. Some of the cyanide/cobalt ratios were very high, as much as 19, but, as thiosulphate was also given, much of the beneficial effect must have been due to that. For rabbits the results were similar, but were also improved by giving thiosulphate. For both species the edetate works adequately up to 2 or 3xLD50 at cyanide/ cobalt ratios greater than unity, as with ordinary cobalt salts, and provided that the dose of dicobalt edetate is not high enough to exert its own toxic effects. There is a suggestion in these findings that the best results are obtained when the dicobalt edetate is given after the hydrocyanic acid, and there was a similar inference to be drawn from all the compounds so far. Hydroxocobalamin as antidote In dosesup to 400 umoles/kg intravenously, no toxic effects of hydroxocobalamin were seen with mice. The substance was rapidly excreted in the urine, which COBALT COMPOUNDS AS CYANIDE ANTIDOTES 463 TABLE 5 DICOBALT EDETATE AS AN ANTIDOTE TO HYDROCYANIC ACID I.p.intraperitoneal; i.v.intravenous; i.m.intramuscular; numbersin'bracketsafterthenumbers ofanimalsdiedgivethenumbers ofanimals tested Hydrocyanic acid Dicobalt Cyanide/ Thio- Numberdied Dose Multiple edetate cobalt sulphate (jmoles/kg) ofLD50 (jtmoles/kg) ratio (g/kg) Total % Mice: hydrocyanic acid (i.p.) followed immediately by dicobalt edetate (Kdlocyanor, i.v.)withand without thiosulphate(i.v.) 240 2'2 25 9-6 0.0 13 (13) 100 240 2'2 50 4'8 00 6(8) 75 240 2'2 100 2'4 00 2(2) 100 240 2-2 240 1-0 00 2(2) 100 480 4.3 25 19X0 0'25 2(7) 29 360 3-2 25 14'5 0 15 0(5) 0 240 2-2 25 9-6 0'4 0(6) 0 240 2'2 50 4'8 0'4 5(6) 84 Mice: pure dicobalt edetate (i.v.)followedimmediately byhydrocyanic acid(i.m.) 600 6-0 75 8'0 0.0 3(3) 100 500 50 75 6-7 0.0 2(2) 100 400 4-0 75 5-4 0.0 6(6) 100 300 3'0 75 4'0 0.0 2(2) 100 200 2-0 75 2'7 0'0 5(13 38 400 4'0 75 5.4 02 4(4) 100 400 4'0 75 5.4 0'5 3 (4) 75 300 30 75 4'0 0'5 4(4) 100 250 2'5 75 3-4 0'25 4(4) 100 200 20 75 2'7 0'25 3(4) 75 200 2'0 75 2-7 0'5 1 (2) 50 Rabbits: dicobalt acetate (Kdlocyanor, i.v.)followedby hydrocyanic acid(i.p.) 220 3'8 240 092 00 1 (1) 100 220 3-8 107 206 00 1 (1) 100 220 3'8 120 1-84 00 1 (2) 50 220 3'8 60 3-7 0.0 1 (2) 50 220 3'8 40 5-5 0-0 1 (1) 100 165 2-8 60 2-75 0'0 4(4) 100 110 1.9 50 2-2 0.0 0(2) 0 67 1.1 50 1-34 0'0 0(2) 0 220 3'8 25 8'7 0'25 1 (3) 33 165 2-8 25 6'6 0-25 0(2) 0 165 2-8 50 3.3 0'5 0(2) 0 Rabbits: hydrocyanic acid(i.p.)followedbypure dicobalt edetate andthiosulphate (i.v.) 165 2'8 25 6'6 0'25 1 (3) 33 165 2-8 50 3'3 0'25 1 (4) 25 became deep red in a few minutes. In these experiments, when mice were used, the theoretical expectations were as a rule approximately borne out. In most of the experiments the hydroxocobalamin was given immediately before the hydro- cyanic acid, this being technically easier, and in these it was found that when the molar ratios of cyanide/cobalt were below unity and the dose of hydrocyanic acid less than 450 pug/kg (about 4x LD50) recovery was the rule, though the results were rather irregular (Table 6). When the hydroxocobalamin was given after the hydrocyanic acid, and after respiration had failed, the results were similar (Table 7), recovery being effected after 500 ttg of hydrocyanic acid (4.5xLD50) provided 2G 464 C. LOVATT EVANS TABLE 6 EFFECTSONMICEOFHYDROXOCOBALAMINFOLLOWEDBY HYDROCYANIC ACID Hydroxocobalamin was given intravenously, and hydrocyanic acid intraperitoneally. Numbers in bracketsafterthenumbersofmicediedgivethenumbers ofmicetested Hydrocyanicacid Cyanide/ Hydroxo- hydroxo- Numberdied Dose Multiple cobalamin cobalamin (Omoles/kg) ofLD50 (pmoles/kg) ratio Total % 600 5*4 600 1.0 4(4) 100 500 4-5 600 0-83 0(7) 0 500 4-5 500 1*0 3 (4) 75 450 40 600 075 2(8) 25 450 4-0 450 1.0 3(4) 75 400 3-6 400 1-0 0(4) 0 400 3-6 600 0-66 0(2) 0 300 2'7 600 0.5 0(4) 0 300 2'7 450 0-66 1 (5) 25 300 2-7 300 1.0 0(3) 0 300 2-7 150 2-0 0(3) 0 200 1 8 225 0-9 0(5) 0 200 1-8 33 6-0 7(7) 100 150 1-4 167 09 0(10) 0 150 14 104 1-45 0(2) 0 150 1-4 87 1-73 0(3) 0 150 1-4 36 4-2 0(2) 0 that an approximately equimolar amount of hydroxocobalamin was given. When smaller quantities of hydroxocobalamin were given, so that, theoretically, one or more LD50 of cyanide was unneutralized, death always occurred. When thio- sulphate was also administered (0.42 g/kg) the results were improved (Table 8). With rabbits, owing to the rather large amounts of hydroxocobalamin involved, fewer experiments were possible; the results were less regular than with mice. Here, although in one experiment 5.2xLD50 was neutralized, most of the animals died at doses over 2xLD50, even when the ratio of cyanide to hydroxocobalamin was unity. Addition of thiosulphate (0.25 g/kg) did not improve the results and, in fact, thiosulphate alone seemed to be more effective than when given with hydroxocobalamin. Thus, after thiosulphate (0.25 g/kg intravenously) followed by 110 Itmoles/kg of hydrocyanic acid intraperitoneally (2xLD50) three out of four rabbits recovered, though after 2.8xLD50 (165 pmoles/kg) three out of three died. We may say, therefore, that thiosulphate alone, administered before the TABLE 7 EFFECTS ON MICE OF HYDROCYANIC ACID AND HYDROXOCOBALAMIN WHEN RESPIRATION HAD FAILED Hydrocyanic acid was given intraperitoneally and hydroxobalamin intravenously. Numbers in bracketsafterthenumbersofmicediedgivethenumbersofmicetested Hydrocyanicacid Cyanide/ Hydroxo- hydroxo- Numberdied Dose Multiple cobalamin cobalamin (pmoles/kg) ofLD50 (Cmoles/kg) ratio Total % 500 4-5 500 1x0 1 (4) 25 500 4.5 400 1-25 0(4) 0 500 4.5 333 1*5 4(4) 100 500 4.5 250 2x0 3(3) 100 400 3-6 600 0-66 1 (4) 25 200 1t8 450 0-45 0(4) 0