The Introduction of Citrate as an Anticoagulant for Transfusion and of Glucose as a Red Cell Preservative

In 1918, Oswald Robertson published a remarkable paper, describing transfusions of stored red cells given near the front line in France in the First World War ( 33). Only 4 years had elapsed since the first few transfusions of citrated blood had been given to human subjects; the method of storage which was used had not previously been applied to transfusion in humans. For the first time, a bank of units of stored blood was created and shown to be of great practical value. Before describing Robertson's background and work, something will be said about the introduction of citrate as an anticoagulant and of glucose as a red cell preservative. These two developments were of vast importance to blood transfusion: the use of citrate made it possible to separate donor and recipient in space; the use of glucose made it possible to separate donation and transfusion by a substantial period of time. 13) noted that CaCl2 had a striking effect on the speed of coagulation as well as on the amount of fibrin formed. Soon afterwards, 6) concluded that calcium salts were absolutely necessary for clotting. They found that if blood is mixed immediately with oxalate (1 g oxalate per 1000 g blood) it remains fluid for several weeks at +3°C and that it clots if calcium salts are added. 29) speculated that the addition of oxalate to blood might be useful in transfusion but would probably be dangerous. However, in the previous year 44) had actually used oxalate successfully as an anticoagulant in transfusion in dogs. He took as much blood as could be drawn off without causing death from haemorrhage, mixed it with oxalate (probably one-tenth volume of 1–2%) to prevent clotting and reinjected it. One of two dogs treated in this way went through the procedure twice in a fortnight. A third dog was transfused with 100 cc of blood from another dog, calculated to be equivalent to one-third of its blood volume. The dogs appeared to be none the worse for their experience. Although 12) stated that ammonium citrate interferes with clotting, he did not understand the mechanism. 30), who found that blood (90 cc) mixed with citrate (10 cc of a 5% solution) remained fluid, realized that the effect was due to the affinity of calcium for citric acid. 5) was puzzled by the fact that anticoagulants such as oxalates and tartrates precipitated calcium whereas citrate did not, and he concluded that its action could not be explained. In a paper written almost immediately afterwards, 39) wrote ‘Je crois, au contraire, qu'il y a une explication possible, et je dirais encore surement démonstrable; la voila.’ He went on to say that citrate, in a dose of about 1.7 g/l of dog's blood, prevents coagulation indefinitely because it reduces the concentration of ionized calcium to below the minimum needed for clotting. Conductivity measurements were used to demonstrate that citrate bound calcium ions, i.e. without precipitation. In the period 1914–15, great excitement was caused in the medical world by the announcement of the first use of citrate for transfusion in clinical practice. No one seems to have pointed out that citrate had been used entirely successfully many years earlier for transfusions in animals. Furthermore, as early as 1893, Wright (later, Sir Almroth) had realized the potential value of citrate in transfusion ( 45). After discussing the fact that citrate keeps blood liquid by decalcifying it, he put in a footnote ‘I may perhaps be allowed to remark in this connection that this variety of decalcified blood (i.e. citrated) appears to be more suitable for purposes of transfusion than even the oxalated blood proposed by me in a previous communication’ ( 44). Of course, as far as humans were concerned, this was an idea too far in advance of its time: there was little point in surmounting the technical problem posed by blood clotting when transfusion remained dangerous because of the risk of incompatibility due to yet-to-be-described natural antibodies. In many animals, in contrast, alloantibodies do not pose a threat in first transfusions and the introduction of anticoagulants had a great deal to offer. In an experiment conducted by 8) at Guy's Hospital, in which rabbits were transfused with allogeneic blood, the citrate method was used ‘because with the artery–vein method of transfusion it is difficult to determine exactly how much has been transfused without intermediate manipulation in vitro’. A total of 0.4–0.5 g of sodium citrate was found to be enough to keep any quantity up to about 100 cc of rabbit blood liquid at 37°C for many hours, provided that blood was obtained ‘from the first main gush which issues from the divided carotid’. In various experiments, about 50–80 cc blood was transfused; the authors commented ‘we had no evidence that the citrate plasma was in any respect toxic.’ The amounts transfused correspond very roughly to giving 2 l of citrated blood to humans. If the results had been known to a clinician interested in human transfusion they would surely have supplied strong justification for trying citrate in man. Another account of the use of citrated blood was provided by 41). Their work was concerned mainly with demonstrating the multiplicity of red cell antigens in cattle, and large amounts of blood were injected intravenously from one animal into another. Animals were bled from the jugular vein into graduated vessels containing a known volume of 4% citrate. The citrated blood from each animal was kept at body temperature and transfused as rapidly as possible into the jugular vein of the other. Eight animals were injected in this way with quantities varying from about 2 to 4 l. In one example, a bull weighing 412 kg was bled of 2.5 l and then received an i.v. injection of 2540 cc citrated blood, equivalent to 2117 cc of pure blood. The animal must then have received 423 cc of 4% citrate, i.e. 16.9 g or about 0.04 g/kg. For comparison, an adult human weighing 70 kg and receiving 450 cc blood mixed with 100 cc of 3% citrate would receive 3 g, which also corresponds to about 0.04 g/kg. Particularly in view of the fact that the transfusions were given ‘as rapidly as possible’, this paper provides further strong evidence of the safety of citrate. In the paper, no reference is made to the previous use of citrated blood for transfusion. 41) seem to have taken it for granted that it was the obvious choice. The present author finds it frustrating to recall that in 1940, when he was introduced to these two scientists over a cup of tea and received a reprint of their work, he did not realize that he had lost forever the opportunity of finding out how they came to use citrate. The first transfusion of citrated blood given to a human seems to have been performed by 15), who wrote that he thought of trying citrate as an anticoagulant because he knew from ‘Gengou's thesis' that citrate had ‘dispersal properties’ and that it could stop a suspension of mastic from gelling. From experiments, he concluded that 0.2% citrate would prevent blood from clotting for 30 min. In the course of his experiments, he noted that red cells to which citrate had been added oxygenated themselves less well than red cells in sodium chloride, and drew the conclusion that citrate should be replaced, at least partially, by a non-toxic substance ‘qui retarderait également’ coagulation; he chose glucose and fixed on a glucose–saline solution. He first transfused citrated dog blood to a dog, then citrated human blood to a dog; finally, in March 1914, he transfused blood from one man to another. Concerned with the question of priority, 16) later commented that one of the accounts which he published in 1914 (he published the identical text in three journals) was listed in the Index Medicus of 14 October 1914 but was not mentioned in the American press. The next to publish was 1). After getting colleagues to test citrate in animals and after receiving a large dose intravenously himself without untoward effects ( 3), on 9 and 14 November 1914 he supervised the transfusion of two patients, each of whom received 300 g of blood mixed with 3 g of sodium citrate. These details were given to La Prensa, the leading daily newspaper in Buenos Aires, and appeared in The New York Herald Tribune on 15 November 1914 ( 1). Two months later, 19) published a short paper reporting that he had found blood mixed with citrate (0.2%) to be non-toxic in dogs, and that he had successfully transfused two patients. He mentioned 15) work and said that Weil had also transfused citrated blood successfully, but he did not mention Agote. 2) believed that his claim to priority was being overlooked, but, as 36) pointed out, Lewisohn had been interested for some years in using citrate in human transfusion and publication of his preliminary results was probably precipitated by the newspaper articles recounting Agote's work. 43) found that guinea pigs or dogs could be practically exsanguinated and then rapidly restored by the transfusion of citrated blood; no unpleasant reactions were observed. Human patients were treated in the same way, using amounts ranging from 10 to 350 cc of citrated blood. Blood was aspirated by syringe from a vein, immediately mixed with citrate (1 cc 10% citrate to 10 cc blood) and injected into the patient using a syringe and stopcock. The blood was fresh in some instances; in others it was 3–5 d old; there was no apparent disturbance of any kind except a transient rise in temperature and polyuria. In 1915, 20) published a second paper in which he described 22 transfusions given to 18 patients. Sometimes, as much as 1000 cc was transfused, but more often 500 cc; 0.2% citrate was found to prevent clotting for 3 d. 20) stated that he could find no reference to previous work on citrate, although he came across 15) paper after he had completed his toxicity experiments in dogs. He pointed out that 43) suggested dose of 1% citrate was practicable only for transfusions not exceeding about 250 cc (untrue!) and that, on the contrary, Hustin, considering that a concentration of 0.2% was too small, had recommended the addition of an equal volume of glucose/saline to citrated blood, rendering his method impractical when, for example, 1000 cc of glucose–saline was added to 1000 cc citrated blood. In conclusion, 20) considered that his own method of using 30 cc of 2% citrate for 300 cc of blood was the best, and this was generally accepted. In retrospect, it seems clear that none of the people involved in the introduction of citrate for human transfusion in 1914–15 realized how much work had already been carried out and that the idea of using citrate for human transfusion had been proposed as early as 1893. One can begin by calculating the amount of citrate which would be required to prevent clotting if 1 mole of citrate bound one atom of calcium firmly. The atomic weight of calcium is 40 and its plasma concentration is about 2.5 mmol/l (≈ 10 mg/dl). The molecular weight of trisodium citrate (dihydrate) is 294, so that a concentration of 2.5 mmol/l would correspond to 73.5 mg/dl plasma or, assuming a packed cell volume (PCV) of 0.4, about 44 mg/dl whole blood. However, citrate does not bind calcium firmly, and 39) considered that a ratio of three molecules of citrate to one of calcium was necessary to prevent clotting. In a discussion of this subject at a much later date, 22) pointed out that with the disodium citrate–glucose solution which they had tested the concentration of citrate in the final mixture of plasma–diluent was well over 20 mmol/l and yet the occurrence of occasional clots was still a problem. The amounts of citrate used by various early authors expressed as approximate molar ratios of citrate to calcium, making the assumptions described above, are as follows: 5:1; 39) 4:1; 8) 10:1; 41) 20:1; 15a) 6:1; 19) 5:1; and 43) 23:1. An early investigation made it clear that citrate is rapidly metabolized. 40) found that 20 s after injecting about 50 mg/kg into dogs and rabbits only about one-third was left in the circulation. In human adults, toxic effects such as prolongation of the Q–T segment of the ECG and even cardiac arrest are observed only when the rate of citrate infusion exceeds about 0.04 mmol/kg/min, corresponding approximately to 1 l of citrated blood in 10 min. In practice, citrate toxicity is a problem only in special circumstances, such as in patients with liver failure (for references and discussion, see 28). In 1915, on the eve of the widespread use of the citrate method, the two best methods of transfusion available were the multiple syringe and cannula method ( 21) and the four-way stopcock method ( 42). With both methods, the blood was transferred so rapidly that it was possible to transfuse large quantities without using anticoagulant. However, donor and recipient had to be in close proximity and at least two trained operators were needed. In writing about blood transfusion for war casualties, L. B. Robertson (a Canadian) mentioned citrate only as a possible method and evidently preferred to use cannulae with multiple syringes, flushed with saline as necessary ( 31). When the same author wrote a second paper 2 years later, he was still cautious. ‘The citrate method is by far the simplest but up to the present time there has not been sufficient evidence at the front to show that it is pre-eminently the method of choice over all others’ ( 32). 34) had no such reservations: whereas other methods required one or two assistants, the citrate method could be performed by a single Robertson used a with a of cc; cc of citrate was used and made on the at and corresponding to 500 cc blood, 1). He that 44 transfusions of cc blood had been given to patients at a 1 the used by Robertson for one through a to a below the of the citrate–glucose and is a a to the another through the and is to a citrate is as far as the after the has been into the vein, the syringe is applied and 500 cc blood were used for 250 cc blood, two of this amount then being taken from a When the blood, the syringe was used to the blood into the and the donor is bled to one of the corresponding to cc blood. Robertson's was For example, that they had used it in 100 at a were bled of using cc of citrate. No were although there were occasional of in the to the of transfusion and by for a to the of the First World a blood that might be of use in concluded that would the of red on methods of considered that, ‘as red cells are almost to of might as They as with various animal red cells and found that glucose and had the effect in on although and were not far The for red cells was found to be glucose was for and human red but the were best kept in solution. of their paper that glucose and had very at 20 d there was a of with glucose but none with They three of human blood with two citrate the of trisodium citrate is but used and of glucose (a mixture later known as solution) and say that the cells remained for 4 In a second paper ( transfusion experiments were were bled of substantial amounts cc or of blood and transfused with a suspension containing an equivalent amount of stored red When the blood had been stored for up to about 14 the concentration was well For example, a total of blood was taken from a rabbit and with an equivalent amount of blood stored in for d. The concentration from to at but was on of three other experiments with red cells stored in were two of storage for d and one of storage for In this it was concluded that all of the transfused red cells left the 2 d of transfusion. no were given of transfusion experiments with red cells stored with although such experiments must surely have been carried out 33). The authors stated that the mixture they had found best was and that red cells in it will in and so that the concentration remains this and for which they did not they recommended that for the of human red cells glucose should be i.e. the solution described in their first paper was a choice as it later that, be used by red The of the red is the of glucose to and and ( of course, in their that red cells are almost to although they are to The only from glucose that can be by are other and and these can also red cell storage ( at a later date, measurements of in were using of was found to be far with glucose than with For example, after storage for 3 weeks with glucose some of the red cells were still present in the 1 after but after storage with for only 2 weeks was to after 1 in although the rate of on storage is far with than with citrate is much the same with the two ( is a little that the of glucose over was not by and but the methods they used for in concentration and red cell were probably too at the time. When the of red cells stored in solution was later with that of red cells stored in the citrate–glucose solution used by transfusion in the period cc 3% trisodium citrate and 10 cc glucose for cc solution was found to be after d at 1 was approximately twice as as with the solution in use ( In solution was almost as as citrate–glucose ( The of solution may be due to its large volume of glucose and to the fact that glucose is solution has one because of its large and citrate the red cells stored with it can be transfused only after a amount of has and a substantial amount of the can be Although the solution was used by Robertson in in the World War it was by citrate–glucose of much volume although only after the of glucose as a red cell had been by red cell measurements ( Robertson was in in but his to 2 years later After to Robertson did his clinical at After he on but in he was by to the he to begin on and he a and with In he in the of which the was He completed work with on in rabbits In the medical of the technical from and Robertson went to work at 3 in near He was evidently with the use of citrate and very after his in France he published a paper describing a method as already was In account of the of blood Robertson was given the for the citrate method to the in some weight as he himself transfusion on a very large in the year or two of the war to the author in When Robertson to blood he chose was then known as the with a of more than because he used the solution recommended by and 500 cc blood was taken into a mixture of cc of glucose and 350 cc of trisodium citrate. The citrate and glucose were to The red cells were found to a as the blood could not be used at the of the red cells had to cc and could then be used because this amount of cell suspension does not more citrate than is given in an transfusion. The was off to and the volume made up to 1000 cc by in The blood cell suspension was through into the giving and was then given 500 cc blood in not less than 10 min. Only i.e. were used to the for ( were from with or from who had from with a of or were although no were carried transfusions were of blood kept for but a few were of blood stored for as as d. to 1000 cc blood was In 22 transfusions were given to 20 patients. were as being to without but yet with a of if given blood. The effects were considered to be as striking as with fresh blood; the of time for which the blood had been stored appeared to have no on its the in who received blood for 3 weeks to be as as in who received blood stored for much Robertson concluded that, 4 weeks is about the for which blood can be kept for transfusion as red cells begin to after Robertson does not seem to have from the of storage but for a period by as with which remains to it After there was a in the of their and and their blood by were not in red cell as the often when fresh blood was used because of of plasma volume. was no evidence of out of the 20 patients were to in concluded that the value of this method in the of a large quantity of blood on for a He mentioned that, in a when of stored blood out, much time was in finding suitable from the as such patients had been on from the whereas or even cc could be taken from well such amounts could not be taken from who were Robertson took stored with and used He was then to another near the this time, he his a with a the the two being with was at one of the He had 20 in when the being in of the he transfusions a the blood he then citrate the of the his packed in were to and near the front ( the time of the in March 1918, Robertson came near to being by the to the but he to the with the in at which time he was the by the In with and he received a from the American of Blood for work with red should to and for of early not available to and for and and for me see their paper about Robertson it was