History

Battle Casualties in Korea: Studies of the Surgical Research Team, Volume II

The Hematologic Response to Wounding and to Resuscitation Accomplished by Large Transfusions of Stored Blood*

A Study of Battle Casualties in Korea

Lieutenant Colonel William H. Crosby, MC, USA
Captain John M. Howard, MC, USAR

This study was undertaken by the Surgical Research Team of the United States Army in order to evaluate present methods of blood transfusion as practiced by American surgical teams at the forward hospitals in Korea. Our military surgeons have become increasingly bold in the use of blood for resuscitation of casualties. Thirty pints of blood and even more may be transfused into a seriously wounded man during the few hours required for his resuscitation and surgery. As a consequence it has been possible to save the lives of many casualties who previously might have died of irreversible shock. Use of such large amounts of blood is, of course, not without its hazards. These hazards would at first glance seem to be compounded by the fact that most of the blood used in the wounded men is collected in the United States and transported by air to Korea. Most of it is 10 to 20 days old before it is used. It is discarded after 21 days. All of the blood is from universal donors. That is to say, it is group O, Rh positive, and is used without cross matching regardless of the blood group of the recipient.

The present study was undertaken to determine what, if any, are the harmful effects of massive transfusions and to learn if the age of the blood and the difference of blood group contribute anything deleterious. Insofar as we were able to tell, they did not. Even after large, rapid transfusions the plasma potassium was rarely elevated and pathogenic amounts of plasma hemoglobin were not encountered. The use of O blood in A recipients undoubtedly resulted in some loss of recipient red cells, as will be shown, but there was no evidence of clinical transfusion reactions.

The hematologic response of these men to transfusion could not be separated from the effects of the wounds that required the transfusion. This report, therefore, is more a study of patients than of transfusions.

*Previously published in Blood, The Journal of Hematology 9: 439, 1954.

Materials and Methods

The technical work was carried out by one of us (WHC) in the laboratory of the Surgical Research Team attached to the 46th Army Surgical Hospital in Korea, from late October 1952 until mid-January 1953. The hospital was located several miles behind the infantry division that it supported. Patients were brought by ambulance or helicopter from the battalion aid stations on the line. They were usually received within several hours after they had been wounded. Often they had received transfusions of whole blood or plasma substitutes at medical stations along the line of evacuation. All of the patients were young, their ages ranging from 18 to 24. The average was about 20.5 years. The group included Caucasians, Negroes, Nisei, Puerto Ricans, and South Koreans. Our study involved 37 patients, each of whom was so severely injured as to require blood transfusion. None of the patients had been burned. The size of the transfusions varied from 2 pints to 42. Eighteen of the patients were given 9 pints or more. Because the Ashby studies were especially valuable, we tended to select patients of other than group O. A battery of hematologic studies was carried out on each patient. In most cases the observations were performed serially and were terminated when the patient`s condition permitted further evacuation. During the period of the study 1,620 pints of blood were used at the hospital. No hemoglobinuric reactions occurred.

The hematocrit of heparinized blood was measured in Wintrobe tubes, filled at the bedside and centrifuged at 2,000 g for 35 minutes. Red cell counts were done by the method of Dacie¹¹ in which 20 cu. mm. of whole blood were diluted in 5 ml. of normal saline (1:250). After thorough mixing the cells were counted in a standard counting chamber. Differential agglutination (the Ashby technic) was carried out by pipeting 0.3ml. of the above red cell suspension into a small test tube containing an estimated 5 mg. of dried anti-A or anti-B serum (Michael Reese). This caused agglutination of the patient`s own red cells but did not agglutinate the transfused group O cells. After 30 minutes the tube was centrifuged, the unagglutinated cells were resuspended and counted in a counting chamber.¹⁷ Baseline inagglutinable counts were not always possible because of transfusions given to patients before arrival at the hospital. Where baselines were possible, the count varied between 7,000 and 15,000 per cu. mm. In making the total red cell counts and Ashby counts at least 1,500 cells were counted; where the two values were almost equal at least 3,000 cells were counted. The Ashby counts were of value to establish the ratio of donor (group O) red cells to native (A, B, or AB) red cells after

transfusion. Serial Ashby counts revealed alterations of the ratio. An increase of agglutinable native red cells suggested a loss of transfused cells. An increase of inagglutinable group O cells suggested a loss of native cells. Under normal conditions transfused cells are replaced by newly generated native cells at the rate of 0.85 per cent per day.¹⁷

White cell counts were performed in the usual way. Platelet counts were carried out by the direct method of Brecher and Cronkite.⁴ The normal range is 200,000 to 300,000 per cu. mm. Reticulocytes were counted on a dried smear of blood that had been mixed with a saline solution of new methylene blue, as described by Brecher.³ Normal reticulocyte counts were usually about 0.6 per cent. The method for plasma hemoglobin, described elsewhere,¹⁰ was modified to use benzidine base rather than benzidine dihydrochloride as the indicator. The results were read on a Coleman Jr. spectrophotometer at 515?. Because an aqueous solution of benzidine was employed these results are probably about 30 per cent low.⁶ Normal plasma hemoglobin is less than 5 mg. per 100 ml. of plasma. The plasma hemoglobin of bank blood was determined from specimens of the blood obtained just as transfusions were started, after the red cells had been resuspended in the plasma.

The plasma of transfused O blood carries anti-A and anti-B antibodies into the recipient. It was of interest to learn if, after large transfusions of O blood, the presence of foreign antibodies could be detected in the plasma of recipients who were A, B, or AB. Cold and warm agglutinin titers were measured in serum or lightly heparinized plasma. There was little or no difference in titer when serum and plasma were tested at the same time. Serum was doubly diluted with saline and mixed with an equal volume of a 2 per cent suspension of red cells in saline to give final dilutions of 1:2, 1:4, etc. The test was carried out against group O cells as well as group A or B. Clumping of the O cells indicated nonspecific agglutination. When the titer was higher against A or B cells than against O cells, the difference was taken to represent the titer of specific anti-A or anti-B agglutinins. The tubes were refrigerated for 1 hour at 5? C., centrifuged for 1 minute at 400 rpm and chilled again for 5 minutes before they were shaken and read. Agglutination was rated from 1 to 4 plus. One plus was small but definite clumps; at 4plus all red cells were clumped in one or two masses. The suspensions were also examined at 37? C. Details of these procedures are described elsewhere.⁸ The direct Coombs test for incomplete adsorbed antibodies was performed according to Dacie`sinstructions.¹¹

Table 1. The Circulating Red Cells after Transfusion of Group O Blood

Patient	Blood Group	Transfusion		Day Postoperative	Total RBC Million/cu. mm.	Inagglutinable RBC Million/cu. mm.	Proportion of Donor RBC % of Total	Hct	MCV cu. ?	Anti-A Cold Agglutinin
Patient	Blood Group	Size (Pints)	Age (Days)	Day Postoperative	Total RBC Million/cu. mm.	Inagglutinable RBC Million/cu. mm.	Proportion of Donor RBC % of Total	Hct	MCV cu. ?	Anti-A Cold Agglutinin
A Multiple lacerations of intestine	A	9	14-16	0 Postop. 1 3 5 10	6.66 5.75 5.25 5.21 5.10	4.27 3.59 3.40 3.47 3.34	64 62 65 67 67	55 48 41 44 44	83 83 78 84 88	Tr at 1:8 0 0 0
P Massive wound of buttocks and lower back, laceration of rectum	A	6	11	0 Preop. 0 Postop. 1 3 6 8	4.70 4.71 4.43 3.67 2.86 3.08	(14,000) 2.41 2.50 1.90 1.23 1.27	0 51 56 52 43 41	39 45 36 30 24 26	83 95 81 82 84 84	0 1:8 1:4 0 0
M Penetrating wound of right chest, diaphragm and liver	A	9	10	0 Preop. 0 Postop. 1 2 3 4	3.60 4.53 4.85 4.40 4.39 4.02	(7,000) 2.68 2.97 2.42 2.34 2.23	0 59 61 55 53 55	29 38 41 34 34 32	81 84 85 77 78 80	0 1:4 1:2 0 0 0
H Penetrating wound of right chest, lung, diaphragm and liver	A	8	11	0 Postop. 1 5 10	4.93 5.77 5.05 5.10	2.13 2.24 1.78 1.64	43 39 35 32	46 46 42 43	93 80 83 84	trace 0 0 0

52-53

Table 1. The Circulating Red Cells after Transfusion of Group OBlood-Continued

Patient	Blood Group	Transfusion		Day Postoperative	Total RBC Million/cu. mm.	Inagglutinable RBC Million/cu. mm.	Proportion of Donor RBC % of Total	Hct	MCV cu. ?	Anti-A Cold Agglutinin
Patient	Blood Group	Size (Pints)	Age (Days)	Day Postoperative	Total RBC Million/cu. mm.	Inagglutinable RBC Million/cu. mm.	Proportion of Donor RBC % of Total	Hct	MCV cu. ?	Anti-A Cold Agglutinin
G Multiple wounds of legs with section of right femoral artery	A	8	17	0 Postop. 1 2 4 6 9 11	4.38 4.34 4.77 4.37 4.24 4.38 4.34	3.39 3.17 3.46 3.15 2.75 2.74 2.53	77 73 73 72 65 63 58	39 39 41 38 35 36 37	89 90 86 87 83 82 85	1:32 1:16 1:4 1:2
I Penetrating wound of abdomen. Died 30 hrs. postop.	A	13	13	0 Aid station 0 Postop. 8 hrs. 1 0 Preop.	5.65 7.01 7.20 5.90 6.50	(11,500) 4.72 4.81 3.93	0 67 67 67	59 65 53 46	84 90 90 71	0 0
K Bilat. fracture of femur	A	7	12-14	0 Postop. 1	5.06 4.71	2.91 2.67	58 57	46 38	91 81	tr 0
II Right chest, arm, thigh	A	6	11	0 Postop. 2 0 Preop.	5.09 4.70 5.37	2.82 2.75	55 58	39 41	83 76
E Multiple wounds including chest	A	8	12	0 Postop. 1 10	5.38 4.64 3.99	3.14 2.69 1.79	58 58 45	47 37 34	87.5 80 85	0
Q Amputation R foot	A	6	13	1	4.53	3.25	72	38	84	0
U Thigh, abdomen	A	4	12	0 Postop. 1	6.4 5.82	1.82 1.57	35 37	51 47	80 81	0 0
V Multiple wounds of extremities	A	3	10	0 Preop. 0 Postop. 1 2 4 5	4.28 4.21 4.85 4.30 3.54 3.41	(5,500) 1.71 1.83 1.50 1.27 1.23	0 24.5 26.5 28.5 28 28	35 38 38 33 30 28	82 90 78 77 85 82	0 0 0 0 0
J Amputation L foot	A	2	13	0 Postop. 1	4.74 4.36	0.90 0.97	19 22	38 38.5	80 88	0 0
N Multiple light wounds	A	2	11	1	4.85	0.98	20			tr at 1:8
O Face and skull	A	2	9-11	1	5.29	1.22	23	40	76	0
R Multiple light wounds	B	2	9	1	5.58	0.96	17	48	86	0
T Multiple light wounds	AB	2	16	1	3.90	0.90	23	30	77	Anti B-O Anti A tr at 1:8

Results

The Bank Blood

The plasma hemoglobin determinations performed on 300 pints of bank blood revealed an average of 34 mg. of hemoglobin per 100 ml. of plasma. In 3 per cent of the bottles the plasma hemoglobin exceeded 100 mg. In 84 per cent the plasma hemoglobin was less than 50mg. As the blood aged, the plasma hemoglobin gradually increased. At 10 days the average was about 25 mg. per 100 ml.; at 20 days it was about 50 mg. In blood carefully drawn into ACD solution at our hospital, the plasma hemoglobin was 5 to 6 mg. The plasma hemoglobin of heparinized normal blood drawn into an oiled syringe is usually 2 to 3 mg.¹⁰

The post-transfusion survival of bank blood was studied by means of the Ashby technic of differential agglutination. The results of some of these studies are presented in Table 1.

The Patients at the Time of Admission

An adequate study was carried out on the blood of 11 patients upon admission to the hospital (Table 2). The leukocyte count was high, usually 20,000 to 25,000 per cu. mm. but in one case it was 40,000 and in another 49,000. Both of these patients had multiple wounds of the legs. In both the evacuation time was long. The platelet count on admission was normal or high. The red cell count was usually about 4 million but the mean corpuscular volume was low and so was the hematocrit. The plasma hemoglobin was normal excepting a slight elevation in one patient who was receiving a transfusion of whole blood at the time of admission.

Table 2. Patients at Time of Admission

Patient	Wound	Evacuation Time (Hours)	Treatment en Route	Hct	MCV cu. ?	Plasma Hb mg./100 ml.	Leukocytes X 1000/ cu. mm.	Platelets Million/cu. mm.	Transfusion (Pints of Blood)
K	Ext	5	0	46	71		23	0.75	7
M	Chest	3	0	29	80	2.5	17	1.25	9
P	Abd	2	1 Gel	39	83	1.6	24		6
Q	Ext	5	2 Dex 2 Bl	31	86	3.8	25		6
R	Ext	2	0	50	90	0.8	19		2
T	Ext	10	2 Alb	31	76	0.5	40	0.8	2
U	Abd	5	1 Dex	46	80		8	0.26	4
V	Ext	2	1 Alb	35	82		27		3
EE	Ext	10	7 Bl	32	82	2.6	49	0.54	9
FF	Chest and Ext	4	1 Bl	32	79	11	22	0.26	9
GG	Abd	2	1 Dex	37	80		23	0.25	13

Abd = abdominal; Ext = extremity; Dex = 500 cc. of .6 per cent Dextran; Bl = 500 cc. of whole blood; Alb = 200 cc. of 25 per cent serum albumin; Gel = 500 cc. of3 per cent modified gelatin; Hct = hematocrit; MCV = mean corpuscular volume. Evacuation time is the time elapsed between wounding and admission to the surgical hospital. The size of the transfusion is stated as an index of the severity of the wound.

Table 3. Leukocytes During Resuscitation and Surgery

Patient	Wound	Hourly Leukocyte Counts (X 1,000 per cu. mm.)
Patient	Wound	Preop.	1	2	3	4	5	6	7
1	Ext; abd	28	27	23	17	21	18	23	16
2	Abd	20	16	15	11	10
3	Abd	4	3	13	8	6	6	3	11
4	Chest; ext	42	49	35	42	34	29	34	39
5	Chest; ext	33	27	26	21	19	15	16
6	Ext; gangrene	19	26	32	13	12	10	14	10
7	Multiple ext	14	14	15	21	22
8	Abd	12	7	11	11	7	9	13
9	Pelvis; ext	25	13	15
10	Femur	13	13	11	9
11	Ext	22	19	18	17	13	14	14
12	Abd	19	24	14	8	8	8	26	31
13	Buttocks; ext	19	11	8	9	10
14	Abd; ext	20	35	43	32	24	21	14
15	Abd; ext	5	5	5	5	6	6	3	2
16	Ext; back	36	69	39	26	29	33	24	20
17	Abd	29	24	11	7	5	5
18	Ext	20	17	18	22	19	22	20
19	Chest; abd	37	17	24	19	13	16
20	Abd	27	37	29	23	26	24

Abd = abdominal; ext = extremity. These patients are of a later series than those of the other tables.

The Patients After Operation

Leukocyte and platelet counts before and after operation are available on 17 patients of this series. In all but two the leukocyte count dropped. The average preoperative count was 24,000 per cu. mm.; postoperative 15,000. The platelet count usually rose although in two patients it fell. The average preoperative count was 450,000 per cu. mm.; postoperative 510,000. These averages do not include patients with hemoclastic reactions.⁹ A second series of leukocyte counts was performed before and during operation at hourly intervals. It demonstrated the various patterns of reaction to injury and transfusion (Table 3). In this second series patients 3, 12, 15, and 17 probably represent hemoclastic reactions characterized by an abrupt disappearance of leukocytes and platelets and a coincident fall of blood pressure. A most severe reaction of this sort was encountered in the patient presented below as case I. At the conclusion of operation his leukocyte count fell to 500 per cu. mm. and his platelet count to

17,000. Five hours later the leukocyte count was 2,700, the platelet count 375,000. At this time clotting time was 90 minutes in silicone, and clot retraction was poor. Eight hours after operation the leukocyte count was 3,510 and at 24 hours it was 11,150. Six different organisms were identified in a blood culture from this patient. In none of the patients with the hemoclastic reaction was fibrinolysis severe enough to cause bleeding.

A mild tendency to ooze from cut surfaces was noted in several patients who received transfusions of banked blood in excess of 20 pints. The bleeding phenomenon was little more than a source of annoyance to the surgeons who were able to control it easily by pressure. In these patients the clotting time, clot retraction, platelet count, and tourniquet test were normal. Prothrombin activity in these patients was studied.²⁰ It was found that the one-stage test of prothrombin time was much prolonged, indicating a defect amounting sometimes to a loss of more than 50 per cent of activity. Further tests indicated that the defect was on a basis of a relative lack of labile factor, one of the adjuvants of prothrombin conversion. It cannot be said that lack of labile factor underlay the mild tendency to bleed. The cause of this phenomenon has not yet been established.

Table 4. Reticulocyte Response to Anemia After Wounding

Patient	Blood Group	Replacement by Transfused Cells (% of total RBC)	Age of Wound (Days)	Hematocrit	Reticulocytes (% of Total RBC)
M	A	60	11	36	1.2
N	A	20	8	32	3.6
P	A	50	8	26	5.1
S	B	65	3	30	4.4
V	A	25	5	27	3.4
X	A	90	7	32	2.0
FF	O	---	9	38	3.2

Excepting FF there is a possibility that reticulocytes may have been destroyed as they appeared by foreign isoantibodies from the transfused group O plasma. The possibility becomes more remote each day after transfusion. In none of these patients was it possible to demonstrate such antibodies in vitro after the first day.

Reticulocyte counts were done on 15 patients. In eight, the reticulocytes were less than 1 per cent before and after operation. In six there were more than 1 per cent but in none more than 2 per cent. In one patient the reticulocytes percentage was 1.6 before operation and 0.6

afterward. Five to ten days after operation reticulocytosis had developed where anemia was present, but the number of reticulocytes was not appropriate to the degree of anemia(Table 4).

The plasma hemoglobin concentration was usually elevated immediately after operation (Table 5). The determination was carried out on 22 patients who had received from 2 to 37 pints of blood. In seven men (who had received 2 to 9 pints of blood) the plasma hemoglobin was normal. It was above 30 mg. per 100 ml. in five cases. The average postoperative hemoglobin level of all 22 patients was 18 mg. With three exceptions, the plasma hemoglobin concentration had returned to normal within 12 hours after operation. In these three, the postoperative level was 40 mg. or higher. In five patients, the plasma hemoglobin was measured frequently enough to establish the rate at which the pigment was cleared from the blood. In patients not in shock, the plasma hemoglobin decreased at rates that varied from 2.2 to 8 mg. per 100 ml. per hour. During severe shock, the rate of clearance was lower: 0.7 to 1.3 mg. per 100 ml. per hour. These results do not include the patients in whom the plasma hemoglobin concentration remained slightly elevated for several days, apparently as a consequence of continuing low-grade hemolysis (Table 6).

Table 5. The Effect of Large Transfusions on the Patients` Plasma Hemoglobin

Patient	Blood Group	Size of Transfusion (Pints)	Age of Blood (Days)	Average plasma Hb of Transfused Blood (mg./100 ml.)	Patient’s Plasma Hemoglobin (mg./100 ml.)		Rate of Decline of Plasma Hb Postop. (mg./100 ml./hour)
Patient	Blood Group	Size of Transfusion (Pints)	Age of Blood (Days)	Average plasma Hb of Transfused Blood (mg./100 ml.)	Preop.	Postop.	Rate of Decline of Plasma Hb Postop. (mg./100 ml./hour)
B	A	20	11	26	47 (after 6 pts. preadmission)	31	5
AA	O	23	18-19	82	2	102	5
F	B	37	11 (26 pts.) 16 (11 pts.)	21 29	25 (after 14 pts.)	40
G	A	8	17	31	12 (after 6 pts.)	8
H	A	8	11	22	3 (after 4 pts. in 6 hrs.)	4
I	A	13	13	NR		58	8 out of shock 0.7 in shock
K	A	7	12, 14	30		10
BB	O	13	11-12	25		15
M	A	9	10	19	2.5	18
V	A	3	10	31		11
P	A	6	11	19	1.6	3
R	B	2	9	19	0.8	1.2
DD	O	15	11	27	0.8	23 (after 11 pts. in 90)	4
Q	A	6	12	19	2.5	7.8’
S	B	6	16	36	3.9 (after 4 pts. in 5 hrs.)	4.6
EE	O	9	15-16	30	2.6 (after 7 pts. preadmission)	2.5
FF	O	9	13	24	6 (after 4 pts.)	8
GG	O	13	16	33		19	2.2
X	A	18	16-19	36		39	2.5
HH	O	0.8	31	132		28	1.5
Y	A	22	14-15	35		18

Table 6. Effect of a Large Transfusion of Group O Blood in a Patient of Group AB

Date	Total RBC (Millions per cu. mm.)	Inagglutinable RBC (Millions per cu. mm.)	Proportion of Donated RBC (% of Total)	Plasma Hemoglobin (mg./100 ml.)	Titer of Anti-A Cold Agglutinin
Nov. 21 8 hrs. postop.	8.55	5.49	64	35	1:128 Direct Coombs test neg.
Nov. 22	7.81	5.26	67	19
Nov. 23	6.68	5.42	81	21	1:128
Nov. 24	6.65	5.52	83	18	1:64
Nov. 25				14
Nov. 26	6.68	5.71	85	5.2	1:64
Nov. 27	6.64	6.13	92	3.9	1:16
Nov. 29	6.81	5.74	84	2.3	1:8
Dec. 1	6.69	5.40	81		0
Dec. 4	7.10	5.41	76		0

Case D. Penetrating wound of the abdomen. 2000 cc. of mixed blood and intestinal contents were removed from peritoneal cavity. The patient was given 14 pints of group O blood in 7 hours, 10 of them within a period of 2.5 hours.

Note the increasing proportion of donor red cells indicating a progressive loss of native red cells. The plasma hemoglobin remained abnormal for several days, probably because of a mild intravascular hemolysis. The specific anti-A agglutinins may have been the mechanism whereby the native red cells were eliminated. Anti-B agglutinins were present in a titer of 1:2 at the time of operation but were gone within an hour postop. There were no nonspecific cold agglutinins until Dec. 1 when they appeared at a titer of 1:4 against group O red cells. The patient secreted B substance in his saliva but did not secrete A substance.

The mild hemolytic phenomenon was not apparent clinically. The patient`s clinical course throughout this entire period was one of continual improvement.

Severe renal insufficiency developed in two patients of this series. One had an abdominal wound. His admission to our hospital had been delayed more than 24 hours. He received 4 pints of blood during his operation and was in moderate shock part of the time. Plasma hemoglobin after operation was 7 mg. per 100 ml. The other patient had an extremely severe wound of the pelvis and lower abdomen. He received 18 pints of blood. His plasma hemoglobin was 39 mg. at the time the last pint was being given. In neither of these men was it possible to relate the development of lower nephron nephrosis to the presence of hemoglobinemia of pathogenic concentration. The highest level of plasma hemoglobin encountered in any patient of this

series was 102 mg. per. 100 ml. in the patient AA. He did not develop renal insufficiency.

No cases of tetany developed on the basis of hypocalcemia from rapid transfusion of citrated blood. The single patient who had tetany did not respond to injections of calcium gluconate but improved when his hyperventilation was controlled.

All blood used for transfusion during this study was group O. The plasma of O blood contains antibodies against red cells of groups A and B. After transfusion, the blood of patients of groups A, B, and AB was examined to find if specific isoantibodies against A and B red cells had persisted. The results of this work are described in detail in another report.⁸ In brief, it was found that the specific cold agglutinins against group A red cells persisted in the plasma of a few patients of groups A and AB. In most cases, the isoagglutinins were rapidly removed from the recipient`s plasma even when he had been given enough blood virtually to replace his own plasma with donor plasma of group O. We found no case in which anti-B agglutinins persisted in the plasma of patients of group B or AB. The Coombs test for incomplete antibodies was also negative in all patients tested. Where the anti-A agglutinin persisted in high titer in the plasms of A or AB patients, there was definite evidence of specific hemolytic activity directed against the native red cells (Table 6). Where the replacement of native red cells by donor red cells was 80 to 90 per cent complete, the slow destruction of native cells continued, even in the absence of demonstrable group-specific antibodies. This hemolytic activity, which was demonstrated by laboratory methods was a low-grade process and had none of the clinical manifestations that are associated with incompatible hemolytic transfusion reactions.

In addition to this specific hemolytic activity there was evidence in severely wounded men of nonspecific hemolysis that destroyed native and transfused red cells indiscriminately. This loss of red cells was computed from blood volume determinations before and after transfusion which failed to account for all of the blood that had been administered.²⁴ It was also demonstrated by serial determinations of blood volume during the first two postoperative days. The loss was not due to hemorrhage. Postoperative bleeding was slight. In all probability, it was not due to sequestration. Some blood volume determinations demonstrated a delayed mixing of the tagged red cells, suggesting sequestration, but this only occurred when there was generalized congestion with a high hematocrit or when there was severe shock.²⁴ Patients B, S, and JJ are examples of the acute hemolytic phenomenon that occurred during the period or resuscitation. In many of the severely wounded there was evidence by cell survival studies of a mod-

erate hemolytic activity during the period of 5 to 15 days after wounding (Table 1).

The more severely wounded men often tended to become anemic during the early days of their convalescence (Table 1). By measurements of red cell mass and plasma volume the falling red cell count and hematocrit were shown to be due to true anemia and not hemodilution.²⁴ Excepting in polycythemic patients an almost linear relationship was found to exist between the red mass and the hematocrit. In general, hemodilution occurred only to the extent that the plasma volume expanded to compensate for the loss of red cells.

Patients who tended to become polycythemic were those with abdominal wounds. In these patients the plasma volume was measured and found to be contracted. It is suspected that plasma is lost into the wounded bowel or into the abdomen. The total blood volume in the polycythemic patients was usually not abnormally great when measured by tagged red cells.²⁴

The two patients who developed acute severe renal insufficiency became rapidly anemic at the time that they began to show clinical evidence of uremia. There had been slight or moderate hemolytic activity prior to this, as shown by changes in the Ashby counts, but with the onset of uremia the loss of red cells was precipitous. In one patient the hematocrit fell from 33 to 23 in 24 hours; in the other the red cell count fell from 5.24 million per cu. mm. to 4.65 million, a loss of 11 per cent in 24 hours. Neither patient had hemoglobinemia or bilirubinemia at the time. Other patients who developed renal insufficiency have been found to be icteric.²⁴

Variations of Mean Corpuscular Volume

Variations of mean corpuscular volume (MCV) were found in all patients who were severely wounded. The MCV was uniformly low at the time of admission (Table 2). Immediately after transfusion the MCV was normal but usually within a day it again became abnormally low (Table 1). Thereafter it gradually returned toward normal. The changes in MCV were not accompanied by corresponding changes in the hemoglobin concentration of the whole blood. In blood smears and in the counting chamber there was no evidence of fragmentation of red cells to account for the reduction of MCV. The MCV of eight normal soldiers living in the same sector of the combat zone as the patients ranged from 86 to 96?.³ In six of the eight the MCV was 90.5 ? 1.0.

Discussion

The appearance of anemia during the days that followed resuscitation depended upon several factors. Red cells were lost in other ways than bleeding and the production of red cells was probably impaired. Except in a general way, it was not possible to define the degree of activity of each of the factors that caused anemia.

The Loss of Red Cells That Had Become Nonviable in the Stored Blood Used for Transfusion

During the refrigerated storage of blood preserved with ACD solution a progressively increasing number of red cells become incapable of survival in the circulation after transfusion. After 2 weeks of storage about 10 per cent of the red cells have become nonviable; at 3 weeks about 25 per cent are nonviable.¹⁷ The Ashby technic has been widely used to study the post-transfusion survival of stored blood. Ashby counts were used in the present study but they showed little evidence of loss of transfused cells during the first 24 hours (Table 1). Indeed, in most cases the proportion of donor cells actually increased during that time though no more blood was transfused. In the interpretation of this observation certain points must be borne in mind. We must presuppose that the loss of nonviable red cells in these patients was at least as great as others have observed in normal recipients examined under better conditions. There are several important differences in the wounded patient:

1. The postoperative Ashby count was not a true reflection of the post-transfusion baseline. As much as 8 hours were consumed in the resuscitation and surgery of seriously wounded men. The nonviable red cells in the first blood given were undoubtedly gone when the postoperative sample was finally taken.

2. During surgery an indeterminate proportion of the donor red cells was lost through bleeding.

3. There was, as will be shown, a specific loss of native red cells when a large amount of group O plasma was transfused into a recipient of another blood group. The specific loss alters the proportion of native to donor red cells in favor of the donor cells, hence, the apparent increase of donor red cells in the first 24 hours after transfusion. An outstanding example of this reaction is shown in Table 6.

4. In some patients who received large transfusions and became polycythemic, an expansion of the circulating blood volume was demonstrated.²⁴ Changes in blood volume make the interpretation of Ashby counts difficult unless the magnitude of the change is known.

5. Severe wounds appear to be associated with nonspecific hemolytic activity against donor and recipient cells alike.

For these reasons, when a massive transfusion was given, it was not possible to compute the proportion of nonviable red cells by the Ashby method. With smaller transfusions these problems did not intrude. A single pint of blood should increase the inagglutinable count of an average-size man by about 0.5 million per cu. mm. The last six patients in Table 1 received small transfusions, and inagglutinable counts indicate little loss of red cells from nonviability during the first 24 hours after transfusion. The studies of plasma hemoglobin (Table 5) demonstrated little evidence of intravascular hemolysis which occurs if stored blood has deteriorated (patient AA). The loss of most nonviable red cells is believed to occur in the reticuloendothelial system without releasing pigment to the plasma.

From all of these data it seems justifiable to conclude that the post-transfusion destruction of red cells due to nonviability was not great. The loss of these red cells, even in association with massive transfusions, did not appear to impose a burden upon the excretory mechanism of the patients.

The Loss of Red Cells Due to Activity of Transfused Isoantibodies

When group O, universal donor blood is transfused into a recipient of another blood group, the plasma carries with it antibodies that are specifically active against the red cells of the recipient. The anti-A and anti-B antibodies consist of warm agglutinins, cold agglutinins, hemolysins, incomplete antibodies, and perhaps others which cannot be demonstrated excepting by their ability to eliminate red cells from the circulation.¹⁷ The incompatibility of transfused group O plasma for the recipient`s red cells is generally believed to be non-pathogenic excepting in the case of the dangerous universal donor. The dangerous donors have such a high titer of pathogenic antibodies that the blood produces acute hemolytic reactions when transfused into group A recipients.¹⁵ No such reaction was observed in this series. During World War II Ebert and Emerson¹³ demonstrated that mild hemolytic activity directed specifically against the recipient`s A or B red cells was not uncommon after transfusions of universal donor blood or pooled plasma. Our own observations confirm this. It was noted above that the Ashby counts usually showed that the proportion of native cells decreased during the 24 hours after transfusion (Table 1). This is presumed to be the effect of transfused isoantibodies. In Table 6 is shown an outstanding example of this sort of reaction. In this patient it was possible to demonstrate a high titer of cold agglutinins specifically active against red cells of

his own blood group. The foreign antibodies left his circulation slowly. It is important to note that the hemolytic activity of these antibodies did not produce any clinical reaction. During the time he was under observation the patient progressively improved. His convalescence was no less rapid than that of others with similar wounds in whom no such hemolysis occurred, they being of group O. However, there is in this situation a matter of great clinical importance. There existed a potential danger to patients who had been given 18 or 20 pints of group O blood, thus virtually replacing their own red cells and plasma by donor cells and plasma.

After several hours or days it often became necessary to give additional blood to such a patient. Fresh blood was available from troops in the neighborhood and there was a temptation to ask for group-specific blood rather than to continue with group O, universal donor blood. The problems of type and cross match were formidable. Ninety-five per cent or more of the red cells in such a patient`s circulation may have been replaced by group O. The identification of his own group on the basis of the agglutination of less than 5 percent of the red cells involved more judgment than one should expect of most technicians. If foreign isoagglutinins persisted in the patient`s plasma, as they did in the patient shown in Table 6, it was impossible to match his blood with donor blood of his hereditary group. It was possible, of course, to obtain a satisfactory cross match if blood were drawn from the patient before the beginning of the transfusion of O blood. It was also possible to give group-specific blood without cross matching. We believe that both of these procedures are dangerous and may lead to several reactions. No such reaction occurred during this study, but in the records of several of the hospitals in Korea we found circumstantial evidence to indicate that such reactions had actually occurred. See the abstract of Case KK. See also page 72. When a man has been given a large amount of universal donor blood, it is recommended that he continue to receive universal donor blood should he need subsequent transfusions during a period of 2 weeks. After this time it seems fairly certain that the foreign isoantibodies will be gone from his circulation.

The Loss of Red Cells Associated With Damage or Destruction of Large Amounts of Tissue

The members of the Surgical Research Team had evolved a rule of thumb: "Abdominal cases become polycythemic; amputees become anemic." The rule was a good one, and its applicability was especially obvious in the severely wounded of each category.

The most obvious difference between the two types of wound was the amount of tissue that was damaged or destroyed. A severe abdominal case may have been wounded by a single small-arms missile that perforated the stomach and colon to cause massive contamination of the peritoneum, often with little loss of blood. This patient would develop a refractory type of shock and was then given 8 to 12 pints of blood, in the expectation that the red cells would fill the relaxed and sluggish parts of his circulation. When his blood pressure recovered, the patient sometimes became plethoric and congested, and sometimes showed signs of incipient pulmonary edema. His hematocrit would lie between 55 and 65. During the period of convalescence when he was under our observation, the hematocrit might recede slowly (case A, Table 1; case D, Table 6; abstract of case AA). The patient who stepped on a land mine and suffered a bilateral high thigh amputation reacted quite differently. With competent tourniquets in place he might be given 12 transfusions in 2 hours and not develop congestion nor a high hematocrit. He might even require more blood to complete his operation and he might need another transfusion the next day or so because of anemia. (Abstracts of cases B, S and JJ are examples of this.) The fate of this transfused blood has not been proved. Certainly it was not all lost by bleeding. Some of it may have been extravasated into wounded tissues. Sequestration was a possibility but not a likely one: these patients may not ever have been in shock, or if they were their blood pressures soon recovered; blood volume studies did not show evidence of sequestration that was sometimes encountered in the plethoric abdominal case and which revealed itself as delayed mixing of injected tagged red cells.²⁴ We are left with the possibility of hemolysis which includes the destruction of extravasated blood. This appears to be a fairly good possibility.

During World War II several investigators^{14, 23} commented upon the tendency of patients with traumatic amputations to become markedly oligemic (a term that implies a reduction of the mass of circulating red cells). Cleghorn et al.,⁵ on the basis of computations of circulating blood volume and size of transfusions given, came to the conclusion that hemolysis must contribute to the loss of red cells. On the basis of similar computations, we have arrived at the same conclusion. Furthermore, it seemed probable that this particular hemolytic process rapidly subsided but did not always disappear. In some patients it persisted well into the period of convalescence. (For examples, see case P, Table 1, or the abstract of case B. Compare them with the abdominal cases, A and H in Table 1.) When the loss of red cells

slowed down, the evidence of persistent hemolysis was apparent in the rapidly declining proportion of donor red cells. This decline has nothing to do with nonviability. The normal rate of replacement of donor red cells is less than 1 per cent a day (patient H). In patient G the donor cell count fell at a rate of 2.5 per cent per day but replacement was adequate and no anemia developed. In patient P it fell at 5 per cent per day, replacement was not adequate, and anemia appeared. In patient S (case abstracts) the loss was at least 12 per cent per day. Several of the patients of this series were evacuated to Walter Reed Army Hospital where the Ashby counts were continued. These counts indicated that after the first few weeks of convalescence the survival of red cells transfused at the time of wounding became normal.

The clinical significance of the distinction between the abdominal wound and the traumatic amputation was thoroughly known to the surgeons who cared for these men. It was important not to push transfusion too vigorously in the patient with an abdominal wound who was not bleeding. It was of equal importance to use blood rapidly and in large amounts in a patient with a bilateral amputation. Such a patient might develop severe anemia while a slow transfusion was actually being administered. The U. S. Army`s Field Research Team in Italy during World War II observed a patient with bilateral traumatic amputation of the thighs, a wound similar to that of our patient, JJ. It was reported that it was not possible to transfuse the patient in Italy rapidly enough to save his life.² The use of air pressure to speed the injection of blood and the use of arterial transfusions may well be lifesaving in such cases. It is recommended that the hemoglobin of these patients be determined accurately, rapidly, and repeatedly during resuscitation and for at least 48 hours thereafter.

The Loss of Red Cells Associated with Acute Renal Insufficiency

Acute nephritis is known to cause anemia. The anemia has been shown to be a result of two factors: abnormal hemolysis and impeded erythropoiesis.^{16, 18} The severely wounded who develop acute post-traumatic renal insufficiency (lower nephron nephrosis) usually become anemic rapidly, as revealed by a falling hematocrit and a shrinking of red cell volume.²⁴ The hematocrit may fall 10 points in 24 hours. This may happen 10 days after wounding when the hemolysis associated therewith has greatly diminished. The loss of red cells in the uremic patient was not accompanied by hemoglobinemia or marked bilirubinemia. The reticulocyte count was somewhat increased in these patients before the onset of uremia, but it fell below 1 per cent even though anemia increased at the same time. This may have been an indication of impairment of red cell production.

It is pointed out above that many of the severely wounded were prone to develop anemia, and that hemolysis is probably a factor in the pathogenesis of the anemia. During the first week or 10 days of convalescence a severely wounded man may destroy his red cells at three to five times the normal rate. Evidence has been published to suggest that the adult human bone marrow can produce hemoglobin at six to eight times the normal rate; blood destruction can therefore take place at six to eight times the normal rate without appearance of anemia.⁷ The bone marrow of the seriously wounded men did not compensate to this extent. When they destroyed their red cells at only four times the normal rate, they became anemic. Reticulocyte counts also showed evidence of impaired response to anemia (Table 4). With an hematocrit of 25 one would expect a reticulocyte count of 15 to 20 per cent instead of 5 per cent.

Pigment Metabolism

If the 300 mg. of hemoglobin from 1 ml. of red cells were released at one time into the plasma of an average-size man (plasma volume 3,000 ml.), it would raise the plasma hemoglobin 10 mg. per 100 ml. If instead of being released into the plasma the same amount of hemoglobin were degraded to bilirubin in the reticuloendothelial system and if all of the bilirubin appeared in the plasma at the same time, it would raise the plasma bilirubin 0.35 mg. per 100 ml.

In the face of a hemolytic process that appeared to involve hundreds of ml. of red cells, as with such patients as B and JJ, it is remarkable that there was little evidence of pigment liberation. The route of its removal remains obscure. Hemoglobinuria did not appear in any patient of this series. Indeed, the level of plasma hemoglobin did not approach the renal threshold (130 mg. per 100 ml.) except in a single case (AA). The slight or moderate elevations of plasma hemoglobin almost invariably subsided to normal within 12 hours. The return to normal was slow in patients with evidence of intravascular hemolysis due to poor quality of transfused blood (AA) or due to destruction of native red cells by transfused isoantibodies (Table 6).

In most of the patients with severe wounds the plasma bilirubin showed a slight or moderate biphasic elevation that was shadowed by a similar increase of urinary urobilinogen excretion.²¹ Patient AA was the only one of this series whose plasma bilirubin exceeded 4 mg. per 100 ml. Patient B was a typical case. Patient JJ with no elevation of plasma bilirubin was exceptional. It has been noted by the Surgical Research Team that patients with paralytic ileus almost

always became jaundiced. The reason for this has not been established. Clinical jaundice was rarely encountered in other patients. The rate of excretion of bile pigment in the feces could not be studied because the severely wounded rarely defecate during the first week or 10 days.

The Leukocytes and Platelets

Leukocytes and platelets of blood preserved in ACD solution are nonviable.¹⁷ Although they may be visible in the stored blood they disappear immediately from the circulation of the recipient. The transfusions given to some patients of this series were large enough virtually to replace their own blood with blood that was, in effect, free of leukocytes and platelets. The leukocyte count was not seriously depressed by large transfusions (Table 3). The production of new leukocytes was adequate to prevent leukopenia. The platelet level was generally unaffected by dilution or replacement of the patient`s blood by large transfusions.²⁰ In fact, the platelet counts usually increased. The production of platelets was adequate and the platelets themselves seemed competent as judged by bleeding time and clot retraction.

These results are somewhat at variance with other observations of the effect of transfusion on leukocyte and platelet levels. Stefanini and Chetterjea²² have reported a slight drop of leukocytes and a marked fall of platelets following the transfusion of small amounts of compatible plasma, even in normal subjects. Desforges¹²has found that exchange transfusion causes a profound fall of platelets in infants treated for hemolytic disease of the newborn and patients given massive transfusions because of gastrointestinal hemorrhage. It is pointed out that the patients in the present series were conditioned by severe injuries incurred several hours before transfusion which evoked leukocytosis and thrombocytosis. This conditioning may have enabled the patients to withstand the large transfusions without developing a lack of leukocytes or platelets.

Changes in the Mean Corpuscular Volume of the Red Cells

The severely wounded showed consistent changes in the mean corpuscular volume of the red cells. The MCV was abnormally small on admission, tended to become normal again during resuscitation and became small again thereafter (Tables 1 and 2). There was no evidence of fragmentation of red cells and the hemoglobin concentration of the whole blood did not change. This indicates that as the cell volume decreased the hemoglobin concentration in the individual cells increased. The changes in volume thus were due to shifts of

water which may in turn have been related to shifts of electrolytes. The problem is being studied further. Thorn and his collaborators²⁷ have described similar changes in red cell hemoglobin concentration following the administration of ACTH and cortisone. The changes of MCV in the severely wounded may reflect a stimulation of the adrenal cortex. Because the change in MCV was sometimes as great as 15 per cent it is suggested that this variable should be taken into account in blood volume studies which involve shock or trauma.

The Hemoclastic Reaction

On several occasions there occurred an abrupt and profound fall of leukocytes and platelets associated with a sudden drop in blood pressure. During the subsequent several hours the counts gradually recovered (Table 3). This was interpreted to have been a manifestation of a hemoclastic reaction, a reaction that involves, in addition to an abrupt disappearance of white cells and platelets, many changes in the globulin enzyme systems of the plasma.⁹ The hemoclastic reaction occurs in response to many types of injurious agents: foreign protein, peptone, incompatible blood transfusion, etc. We observed the reaction in men who had severe abdominal wounds with gross soiling of the peritoneum. It was suspected that the reaction may have been due to contamination of the blood stream via the lymphatics of the diaphragm. During the hemoclastic reaction in one man a blood culture revealed the presence of six different organisms.

Effects of Massive Transfusions

Evidence of the deleterious effects of massive transfusions did not materialize in the course of our investigation. Pulmonary congestion appeared in several patients who were given large rapid transfusions, but full-blown left-sided heart failure was not encountered. It is significant that these patients were healthy before they were wounded, and they were 20 years of age. Fibrinolysis did not occur as a clinical hemorrhagic state nor could it be demonstrated in incubated clots. This is at variance with civilian experience with massive transfusions during operations for widespread cancer; mild degrees of fibrinolysis are commonplace and severe hemorrhagic reactions have occurred from time to time.^{*, 25}

*Subsequent to this study it is probable that several fibrinolytic reactions were encountered in the forward surgical hospitals in Korea. They occurred in the course of massive transfusions and were characterized by excessive bleeding and by apparent incoagulability of shed blood in vitro. There were no fatalities among the patients of whom we have knowledge. The rate of fibrinolytic reac-

The plasma hemoglobin of the bank blood was not unduly high, an indication that the blood had been well preserved during shipment. The levels of free hemoglobin would have been much higher if refrigeration had not been adequate. Following large transfusions of stored blood the plasma hemoglobin of the recipients was not especially high. It was usually of the same order as that of the plasma of the bank blood given them (Table 5), and that hemoglobin rapidly disappeared. Even though 10 to 20 per cent of the red cells may have been nonviable their hemoglobin was not released into the blood stream where it would impose an excretory burden on the kidneys. Among the casualties in Korea it has not been possible to relate the incidence of post-traumatic renal insufficiency with hemoglobinemia. It had been shown¹⁹ that the combination of prolonged deep shock and the infusion of methemoglobin solutions in dogs can produce a renal tubular lesion similar to that of post-traumatic renal insufficiency (lower nephron nephrosis).The concentration of plasma hemochromagen that occurred in these dogs was 800 to 2,000 mg. per 100 ml. This severe hemoglobinemia caused no demonstrable lesion if hypotension was inadequate or brief.

The post-transfusion hemoglobinemia encountered in Korea has not approached these concentrations except where an obvious incompatible transfusion reaction occurred, and these have been rare. The Renal Treatment Center of the U. S. Army receives all patients of the United Nations Forces who develop renal insufficiency. In 1952 only four patients were transferred to that Center with evidence of post-transfusion hemoglobinuria.²⁶ Over 60,000 transfusions were given in Korea in 1952. This helps to place in proper perspective the role of free hemoglobin in the pathogenesis of post-traumatic renal insufficiency.

The use of universal donor blood in the combat zone is a valuable military expedient. Because only one blood group is involved, an adequate reserve can be maintained with fewer units of blood. Small but effective banks of blood can be established at the forward aid stations. By eliminating the need for a cross match, transfusion can be started as soon as a patient is admitted. It also eliminates the incompatible transfusion reactions that inevitably occur where group-specific blood is used. That only four cases of post-transfusion hemoglobinuria were admitted to the Renal Treatment Center in 1952 indicates that incompatible transfusion reactions are rare under this system. It is of interest that each of these four cases resulted from

tions based on this information is estimated to have been less than 1 per cent among patients who received more than 10 pints of blood. (Personal communications from Major C.P. Artz, MC, U. S. Army, and Commander R. N. Grant, MC, U. S. Navy.)

the use of group-specific blood that was locally procured.²⁶ In two cases there apparently was a mistake in typing or cross matching. The other two were similar to patient KK whose case is abstracted below. Field conditions being what they are, it is not surprising that occasional errors have occurred. It was fortunate that there was little need for local procurement of blood. The blood that was sent to Korea was obviously correctly typed and the dangerous universal donors were adequately identified.¹

Summary

1. During the third winter campaign in Korea, the hematologic reaction to wounding was studied in 37 casualties at the time of initial resuscitation.

2. Of particular interest was the effect of massive transfusions of stored blood. The results of storage of blood-high plasma hemoglobin and potassium, low labile factor activity, nonviable platelets and leukocytes-had little deleterious effect on patients who received as much as 20 to 30 pints in less than 6 hours. The loss of transfused red cells because of nonviability was no more than expected.

3. At the time of resuscitation and shortly thereafter, there was a remarkable loss of circulating red cell mass in patients with wounds that involved much tissue destruction. It is believed that the loss was due to hemolysis but the mechanism is unknown. The loss of red cells may be so rapid that a patient with bilateral traumatic amputation of the legs and an adequate hemostasis would become severely anemic if one hesitated to use large, rapid transfusions. Patients with severe shock whose wounds involved less tissue damage (lacerated colon, for example) did not destroy red cells in this fashion. After moderate transfusions, such patients often became polycythemic. Transfusions had to be carried out rather gingerly because of a tendency to develop signs of congestion.

4. During the early days of recuperation from severe wounds the patients often tended to become anemic. The anemia appeared to be the result of hemolytic processes plus a relative inhibition of red cell formation.

5. Universal donor blood, group O, was used in all transfusions. In patients who are not group O the massive transfusions resulted in the virtual replacement by red cells of another group. The patient`s plasma sometimes contained antibodies against red cells of his hereditary blood group. Gradual hemolysis of native red cells by transfused antibodies was observed. This was not a clinical hemolytic reaction and did not appear to be detrimental to the patient. The presence of the foreign antibodies made it impossible in some cases to cross

match the patient with blood of his hereditary group and suggested a source of danger in attempting such transfusions. After transfusion with the universal donor blood has begun, it is recommended that no change be made to blood of another group until at least 2 weeks have elapsed.

6. No incompatible transfusion reactions were encountered. Several hemoclastic reactions may have been caused by gross contamination of the blood stream from the site of wounds or from the peritoneal cavity.

Abstracts of Representative Cases*

Case A: Age 20, weight 145, blood group A, was wounded in action 5:00 p. m., November 13, with multiple penetrating wounds of the abdomen. He was admitted to the hospital at 9:00 p. m. in a condition of moderate shock. He was found to have multiple lacerations of the descending colon and was bleeding from the mesentery. The left femoral and brachial arteries were perforated. Both were repaired. During resuscitation and surgery the patient was given 9 pints of whole blood. At the conclusion of surgery, his hematocrit was 55. Ten days later without having been given any further transfusion, the hematocrit was 44. Throughout this period the proportion of group O red cells in the circulation remained 65 per cent.

Case AA: Age 22, weight 165, blood group O, was seriously wounded in action 9:30 a. m., November 17, by a mortar fragment which penetrated the left side of his chest, left lung, diaphragm, stomach, spleen, pancreas, left kidney, and lodged in the psoas muscle. He was in shock on admission to the hospital, was resuscitated and again went into shock during surgery. During this time he was given 23 pints of blood. His left kidney was removed. Twenty-one of the 23 pints of blood were 18 or 19 days old, 5 pints had a plasma hemoglobin of 100 ml. or more. The average plasma hemoglobin of the 23 pints was 82 mg. per 100 ml. which was more than double the average of all of the pints of blood examined during this study. These bottles were all of the same lot, which suggested that a break in refrigeration had occurred. Immediately after operation (9:00 p. m.) the patient`s plasma hemoglobin was 102 mg. per 100 ml., plasma bilirubin 2.1 mg. The next morning the plasma hemoglobin was 24 mg. per 100 ml., the plasma bilirubin 8.1 mg., 3.8mg. of which was indirect. The patient never became oliguric. His NPN on the third postoperative day was 67 mg. per 100 ml. The patient`s hematocrit receded slowly. It was 60 on the first postoperative day. Fifteen days later it was 48.

Case B: Age 19, weight 140, blood group A, was seriously wounded in action 5:00 p. m., November 15, when he stepped on a land mine. The injury necessitated high amputation of both thighs. He was admitted to the hospital at 9:00 p. m. having received 6 pints of blood en route. The tourniquets on his thighs were competent. Donor red cells represented 63 per cent of the total in his circulation. The hematocrit was 40, total blood volume (Ashby) 4,000 ml., plasma hemoglobin 34 mg. per 100 ml., plasma bilirubin 0.4 mg. per 100 ml. He was given 6 more pints of blood before operation. During surgery he was given 8 more pints of blood. His blood loss did not exceed 1,500 ml. After surgery the donor red cells were 87 per cent of the total. The hematocrit was 39;

*Details may be found in the tables.

plasma hemoglobin 31 mg; plasma bilirubin 2.3 mg. Six hours later the donor red cells were 98 per cent of the total.

He had been given no transfusion meanwhile. Presumably many of his native red cells had been eliminated by the activity of anti-A antibodies in the transfused plasma. At this time his plasma hemoglobin was 2 mg., plasma bilirubin 2.6 mg., hematocrit 36. The blood volume was 3,500 ml. (Evans` Blue). The patient was not in shock any time after transfusion was started at the battalion aid station. Following operation the volume of red cells to be accounted for was as follows: 1,500 ml. in his circulation at the time of admission plus 2,600 cc. in the 14 pints of blood transfused thereafter. It was possible to account for 1,300 ml. of red cells which were in the patient`s circulation after operation, plus 1,000 ml. which had been lost during operation and by specific hemolysis of native red cells. The discrepancy amounts to 1,800 ml. of red cells, which is the equivalent of about 10 pints of bank blood.

Case DD: Age 19, weight 160, blood group O, was accidently wounded in the shoulder with a small-arms missile at 3:00 p. m., December 17. He was evacuated to the hospital by helicopter arriving at 4:30 p. m. He had a massive hemothorax of the left side of the chest. He was given 4 pints of blood before operation and taken to the operating room at 6:30 p. m. When his chest was opened there was a sudden severe hemorrhage from a laceration of his left subclavian vein. Between 7:00 and 8:00 p. m. he was given 11 pints of blood. At the conclusion of this rapid transfusion, his plasma potassium concentration was 4.65 mEq. per liter. His plasma hemoglobin was 23 mg. per 100ml. His white cell count was 20,000 per cu. mm. At 11:00 p. m., the plasma potassium was 4.17 mEq. and the plasma hemoglobin 11.5 mg. The plasma hemoglobin of the transfused blood averaged 23 mg. per 100 ml. The blood was 11 to 14 days old. The next morning the patient`s hematocrit was 47. His red cell volume (radiochromium) was 1,550 ml.

Case I: Age 20, weight 170, blood group A, was seriously wounded in action, 9:00 p. m., December 1, by a shell fragment that penetrated the transverse colon and proximal jejunum. He was admitted to the hospital at 3:00 a. m. the next morning. He was given 4 pints of blood preparatory to operation. He went to the operating room at 6:00 a. m., was given 8 more pints of blood by 8:00 a. m. His blood pressure fell throughout surgery. Two liters of mixed blood and intestinal contents were removed from his abdomen. Otherwise blood loss was slight. At the conclusion of surgery the patient had signs of pulmonary congestion and appeared plethoric. His hematocrit at this time was 59. During the next hour the hematocrit rose to 64; blood volume determination (red cells tagged with radioactive chromium) showed the red cell volume to be 7,600 ml. Therefore, the total blood volume was approximately 10,200 ml. This patient had a severe hemoclastic reaction at the conclusion of surgery. He died 30 hours after operation.

Case JJ: Age 21, weight 170, blood group O, was severely injured while hunting when he stepped on a land mine, causing a high bilateral traumatic amputation of his thighs. The accident was witnessed by two medical soldiers from our hospital who applied tourniquets, stopped an ambulance and brought him to the hospital about 45 minutes after the accident occurred. He was exsanguinated, pulseless; respiratory rate 8 per minute. Intravenous transfusions were commenced on both arms and the injection of blood was speeded by pumping air into the bottles. A surgeon cut down on the stump of the left femoral artery and when a cannula had been inserted a rapid intra-arterial transfusion was begun. Within 30 minutes he had received almost 20 pints of blood. He recovered consciousness

and blood pressure, but developed transient pulmonary edema. The transfusion was slowed.

On the way to the operating room the tourniquets became dislodged and before they could be reapplied he lost 2 liters of blood. The transfusion was continued. Upon induction of anesthesia the patient again fell into shock. The anesthetic was stopped, and the patient again recovered. He was then returned to the ward without surgery. The tourniquets had remained competent and subsequent blood loss was negligible. The period of resuscitation had occupied 5 hours during which the patient received 32 pints of bank blood. At the time of his return to the ward the hematocrit was 39. The next morning it was 38, and he was given 4 pints of fresh, locally procured blood. At this time the plasma hemoglobin was 2 mg. per 100 ml. and the plasma bilirubin was 0.7 mg. A day later the hematocrit was 28, and there had been no bleeding except for some oozing from the stumps of his thighs. He was given 6 more pints of fresh blood and the stumps were amputated. The patient recovered. He did not become oliguric.

Case KK: Age 19, blood group B, was seriously wounded in action 9:00 p. m., July 22, by an enemy mine. There were severe wounds of the head, abdomen, both legs, and left arm. He was admitted to the hospital at 5:30 a. m. having received 2 pints of blood en route. He was given 4 more pints and taken to the operating room at 8:00 where laparotomy, craniotomy, and d?bridement of the extremities were carried out. When he left the operating room at 12:00 noon he had received a total of 17 pints of group O blood 13 to 15 days old. His plasma hemoglobin was 55 mg. per 100 ml.; plasma potassium was 5.8 mEq. per liter; hematocrit 43. That evening because of unsteady blood pressure the transfusion was resumed. The patient was given 6 pints of fresh group B blood. During this transfusion the patient`s blood pressure fell even further. His plasma was found to be dark red and contained 990 mg. of hemoglobin per 100 ml. The plasma potassium at this time was 6.4 mEq. per liter. During the night the patient`s blood pressure was maintained above 100 mm. Hg by norepinephrine, but it fell the next morning and the patient died at noon. After his operation he produced 485 ml. of urine but none after hemoglobinemia appeared.

This patient was not one of our series. The information was obtained from hospital records.

Case S: Age 22, weight 135, blood group B, was wounded in action 1:30 a. m., December 22, with multiple wounds of both legs, necessitating mid-thigh amputation of the left leg. He was given 3 pints of blood at the battalion aid station and en route. He arrived at the hospital at 7:00 a. m. He was never in shock. On the basis of Ashby counts performed before operation it was estimated that 50 per cent of his blood had been replaced and that his blood volume was approximately normal. Hematocrit was 46. He had lost by bleeding and hemolysis perhaps 1,000 ml. of his own red cells. Loss during surgery was less than 400 ml. of red cells. During evacuation and surgery he was given a total of 6 pints of whole blood, which is the equivalent of about 1,100 ml. of red cells. (The hematocrit of bank blood is about 37.) We assume that the patient had 2,000 ml. of red cells before he was wounded. He received by transfusion 1,100 ml. of which 20 percent, at the outside, were lost by reason of nonviability. Therefore, 2,900 ml. must be accounted for. Before the operation 1,000 ml. had been lost; during the operation 400 ml. were lost; 2,900 ml. less 1,400 ml. would leave 1,500 ml. to be accounted for in his circulation after operation. On December 24, his red cell volume (radiochromium) was 700ml., a discrepancy of 800 ml. The plasma volume (Evans Blue) was 2,460 ml. Hematocrit was 30. The decline in the Ashby count between December 23 and 24 indicated that the transfused cells

were being replaced at a rate of 12 per cent per day, at a time when most or all nonviable red cells would have been removed. Such a change in the Ashby count is characteristic of hemolytic disease where the bone marrow produces red cells for replacement. The patient`s reticulocyte count on December 24 was 4.4 per cent.

References

1. Akeroyd, J. H., Hollingsworth, J. W., and Crosby, W. H.: An Evaluation of the "Universal Donor." In preparation.

2. Board for the Study of the Severely Wounded: The Physiologic Effects of Wounds, p. 56. U. S. Government Printing Office, Washington, D. C., 1952.

3. Brecher, G.: New Methylene Blue as a Reticulocyte Stain. Am. J. Clin. Path. 19: 895, 1949.

4. ----- and Cronkite, E. P.: Morphology and Enumeration of Human Blood Platelets. J. Appl. Physiol. 3: 365, 1950.

5. Cleghorn, R. A., Chute, A. L., and Lathe, G. H.: Studies of Bodily Reactions to Injury in Battle Casualties. III. Observations Following Operation and Typical Histories of Cases with Wounds of Extremities and Trunk not Involving Thorax or Abdomen. Proceedings of the Eighth Meeting of the Associate Committee on Army Medical Research, National Research Council of Canada, Vol. 2, December 1945.

6. Creditor, M. C.: The Quantitative Determination of Plasma Hemoglobin by the Benzidine Reaction. J. Lab. & Clin. Med. 41: 307, 1953.

7. Crosby, W. H., and Akeroyd, J. H.: The Limit of Hemoglobin Synthesis in Hereditary Hemolytic Anemia. Am. J. Med. 13: 273, 1952.

8. ----- and -----: Some Immunohematologic Results of Large Transfusions of Group O Blood in Recipients of Other Blood Groups. A Study of Battle Casualties in Korea. Blood 9: 103, 1954. (Chapter 7, this volume.)

9. ----- and Staffanini, M.: Pathogenesis of the Plasma Transfusion Reaction with Especial Reference to the Blood Coagulation System. J. Lab. & Clin. Med.40: 374, 1952.

10. ------ and Dameshek, W.: The Significance of Hemoglobinemia and Associated Hemosiderinuria, with Particular Reference to Various Types of Hemolytic Anemia. J. Lab. & Clin. Med. 38: 829, 1951.

11. Dacie, J. V.: Practical Haematology. Chemical Publishing Company, Brooklyn, 1951.

12. Desforges, J. F.: Personal communication.

13. Ebert, R. V., and Emerson, C. P.: A Clinical Study of Transfusion Reactions: the Hemolytic Effect of Group-O Blood and Pooled Plasma Containing Incompatible Isoagglutinins. J. Clin. Investigation 25: 627, 1946.

14. Emerson, C. P., and Ebert, R. V.: A Study of Shock in Battle Casualties. Ann. Surg. 122: 745, 1945.

15. Ervin, D. M., Christian, R. M., and Young, L. E.: Dangerous Universal Donors. Blood 5: 553, 1950.

16. Loge, J. P., Lange, R. D., and Moore, C. V.: Characterization of the Anemia of Chronic Renal Insufficiency. J. Clin. Investigation 29: 830, 1950.

17. Mollison, P. L.: Blood Transfusion in Clinical Medicine. Charles C. Thomas, Springfield, Illinois, 1951.

18. Muirhead, E. E., Jones, F., and Grollman, A.: The Anemia of Renal Insufficiency as Induced by Bilateral Nephrectomy of the Rabbit: with Emphasis on its Hemolytic Nature. J. Lab. & Clin. Med. 39: 505, 1952.

19. Rosoff, C. B., and Walter, C. W.: The Controlled Laboratory Production of Hemoglobinuric Nephrosis. Ann. Surg. 135: 324, 1952.

20. Scott, R., and Crosby, W. H.: Changes in the Coagulation Mechanism Following Wounding and Resuscitation. Blood. 9: 609, 1954. (Chapter 8, this volume.)

21. -----, Howard, J. M., and Olney, J. O.: Urine Urobilinogen Excretion following Wounding and Resuscitation. A Preliminary Report from the Surgical Research Team, U. S. Army in Korea, February 1952.

22. Stefanini, M., and Chatterjea, J. B.: Studies on Platelets, IV. A Thrombocytopenic Factor in Normal Human Blood, Plasma, or Serum. Proc. Soc. Exper. Biol.& Med. 79: 723, 1952.

23. Stewart, J. D., and Warner, F.: Observations on the Severely Wounded in Forward Field Hospitals with Special Reference to Wound Shock. Ann. Surg. 122:129, 1945.

24. Surgical Research Team of U. S. Army. Unpublished data.

25. Tagnon, H. J.: Personal communication.

26. Teschan, P. E.: Personal communication.

27. Thorn, G. W., et al.: Pharmacologic Aspects of Adrenocortical Steroids and ACTH in Man. New England J. Med. 248: 284, 1953.