Intraoperative Blood Salvage in Penetrating Abdominal Trauma: a Randomised, Controlled Trial
- First Online:
- Cite this article as:
- Bowley, D.M., Barker, P. & Boffard, K.D. World J. Surg. (2006) 30: 1074. doi:10.1007/s00268-005-0466-2
- 260 Views
Blood is a scarce and costly resource. Transfusion is often required after major trauma but blood may not be readily available, and concerns remain over the potential adverse consequences of allogeneic blood transfusion. Intraoperative blood salvage (IBS) is used extensively after blunt abdominal trauma, but when blood is contaminated by enteric contents its use has been considered contraindicated.
This was a randomised, controlled trial conducted with Ethical Review Board approval in an inner city trauma unit. Patients with penetrating torso injury requiring a laparotomy and who had exhibited hypotension either pre-hospital or on arrival and in whom there was considered to be significant blood loss were randomly assigned to 2 groups. The control group was given allogeneic blood transfusion at the discretion of the attending medical staff. The cell save (CS) group underwent IBS with transfusion of both allogeneic and autologous blood. All patients received prophylactic antibiotics. The primary outcome was exposure to allogeneic blood up to the first 24 hours postinjury.
Twenty-three patients were entered into the control group and 21 into the CS group. The groups were equivalent in demographic details, injury patterns and injury severity. The mean volume of salvaged blood re-infused in the CS group was 1,493 ml (range 0–2,690 ml). The mean number of units of allogeneic blood transfused in the first 24 hours in the control group was 11.17 compared to 6.47 in the CS group (P = 0.008). Enteric injury had been sustained in 17/23 (75%) of the control group and 18/21 (85%) of the CS group (P = NS). Survival in the control group was 8/23 (35%) compared to 7/21 (33.3%) in the CS arm (P = NS). Patients with documented postoperative sepsis were significantly more likely to die compared to those without sepsis (P = 0.04); however, those patients in the CS arm were no more likely to develop sepsis compared to those who received allogeneic blood alone.
In this randomised, controlled trial for patients with penetrating abdominal injuries, IBS led to a significant reduction in allogeneic blood usage with no discernable effect on rates of postoperative infection or mortality.
Uncontrolled bleeding is a major cause of death after trauma, and there is a correlation between transfusion of blood products and morbidity.1 Approximately 40% of the 11 million units of blood transfused in the United States each year are utilised for emergency resuscitation.2 The demand for blood is increasing, but the population of qualified, willing and healthy donors is in decline. In addition to the possible hazards of transfusion, logistic difficulties concerning the safe supply of blood maybe particularly severe during military conflicts or in other austere environments. According to the International Committee of the Red Cross, the availability of blood for transfusion in these settings is “a major determinant of outcome”.3
Intraoperative blood salvage (IBS) has the potential to diminish the volume of allogeneic blood necessary to replace massive haemorrhage during emergency surgery. IBS is used extensively after blunt abdominal trauma,4 but when blood is contaminated by enteric contents, its use has been considered contraindicated.5,6 As clear precedent exists for reinfusion of culture-positive cell-saver blood,7–11 we designed a trial to establish the feasibility of using IBS after major penetrating abdominal trauma.
This was a randomised, controlled trial conducted with Ethical Review Board approval at the Johannesburg Hospital Trauma Unit, part of the Department of Surgery of the University of the Witwatersrand, Johannesburg, South Africa.
Assuming the cell saver would result in a 40% reduction in blood requirement with standard deviation of 4.5 units, for a power of 80% and type 1 error rate of 5%, it was calculated that there would need to be 20 patients in each arm of the study.
Patients were assessed on arrival in the Emergency Room by a single investigator and enrolled in the study if they had penetrating torso injury requiring a laparotomy and had exhibited hypotension (< 90 mmHg) either pre-hospital or on arrival and in whom there was considered to be significant blood loss. Patients under the age of 18 years and those patients whose injuries were greater than 6 hours old were excluded.
Consent was obtained preoperatively from the patient when possible or from a family member if present. However, in accordance with other studies involving severely injured trauma patients,12 ethical approval was granted for consent to be waived if it was not possible to obtain full consent prior to laparotomy.
Randomisation was by opening of envelopes containing dedicated data collection sheets previously assigned to either group by computer-generated random number tables. The control group were given allogeneic blood transfusion at the discretion of the attending medical staff. The cell save (CS) group underwent IBS using a Cell Saver 4 machine (Haemonetics, Braintree, MA, USA) with transfusion of both autologous and allogeneic blood as required.
Intraoperative blood salvage entails the collection and reinfusion of blood lost during surgery. Shed blood is aspirated from the operative field into a reservoir via a double-lumen suction catheter incorporating an anticoagulation line with heparin being added at the tip of the suction apparatus. The reservoir is designed to filter out debris and clots greater than 150 μm in diameter. Before blood collection, the reservoir is primed with 200 mL of heparin saline solution (30 U/mL). Further, heparin saline solution runs through the system to the tip of the suction catheter and then back through the sucker into the reservoir. The cell saver process removes the heparin to insignificant levels during the wash cycle. The surgeons attempted to avoid suctioning blood that was grossly contaminated with faeces or other debris within the operative field and blood air interface was avoided when possible during aspiration to minimise fragmentation of red cells. Soaked laparotomy pads were placed in a bowl with a small quantity of warmed saline, gently agitated and the blood solution then aspirated. The aspirated and anticoagulation blood was macro-filtered and then pumped into a centrifuge bowl. This process separates the blood into its two major components, packed red blood cells and plasma. The packed red blood cells are then washed with normal saline. The used washing solution containing saline, plasma proteins, free haemoglobin, cellular stroma, platelets, leucocytes, activated clotting factors, anticoagulants, and other contaminants or debris is discarded.
The washed packed red blood cells are pumped from the centrifuge bowl into a transfer pack and then transfused to the patient through standard blood administration units. The machine was used in automatic mode; therefore the average haematocrit of the salvaged cells was 55%, meaning that 225 mL of processed blood is equivalent to 1 unit of homologous red cells. Although the oxygen transport properties and survival of red cells are similar to bank blood, salvaged blood does not clot normally as coagulation factors and platelets are absent.5 All patients received prophylactic antibiotics, according to the preferred schedule of the operating surgeon. In the IBS group, a sample of the processed blood was taken just before reinfusion, for microbiological examination. An independent adjudicator assigned Injury Severity Score (ISS), Revised Trauma Score (RTS) to the patients. Penetrating Abdominal Injury Severity Score (PATI) was assigned by the operating surgeon immediately after the operative procedure; patients with a PATI score > 25 are considered to be at high risk of infectious morbidity.13 The primary outcome for this study was exposure to allogeneic blood up to the first 24 hours post injury. Statistical analysis was with the unpaired t test and Fisher’s exact test, using Graphpad Software (San Diego, CA, USA).
Patient characteristics by study treatment arm
Median pre-hospital time (when known)
Median ER to OR time
Revised Trauma Score
Injury Severity Score
Penetrating Injury Severity Score
Number of organs injured
3 (±SD 1.4)
The mean volume of salvaged blood retransfused in the CS group was 1,493 mL (±SD 617.43 mL, range 0–2,690 mL). The mean number of units of allogeneic blood transfused up to the first 24 hours postinjury in the control group was 11.17 (±SD 6.06) compared to 6.47 (±SD 5.14) in the CS group (P = 0.008). Survival in the control group was 8/23 (35%) compared to 7/21 (33.3%) in the CS arm (P = NS). For patients with enteric injury in the control arm, the survival rate was 4/17 (23.5%); for CS patients with enteric injury, the survival rate was 7/18 (38.8%). This is not significantly different between the 2 groups (P = 0.47).
Patient outcomes by study treatment arm
Bank blood transfusion in first 24 hours
11.17 (±SD 6.06)
6.47 (±SD 5.4)
Cause of death
Initial culture of salvaged blood and subsequent postoperative blood cultures in cell-saved patients
Initial blood culture
Postoperative blood culture
Coagulase-negative Staphylococci, Citrobacter diversus, Acinetobacter, Enterobacter
Escherichia coli, Morganella morganii, Enterococcus faecium, yeast
Escherichia coli, Alcaligenes faecalis, Clostridum species
Enterobacter, Pseudomonas, Coagulase-negative Staphylococci, Klebsiella
Escherichia coli, Enterococcus faecium, Klebsiella
Control patients who did not die from exsanguination, their postoperative blood cultures and survival
Control patients’ postoperative blood cultures
Coagulase-negative Staphylococci, Enterobacter
Escherichia coli, Bacillus cereus, Group B Streptococcus, Staphylococci aureus, Acinetobacter
Escherichia coli, Proteus
The mean volume of Fresh Frozen Plasma in the control group in the first 24 hours was 4.04 (±SD 4.3) compared to 4.76 (±SD 4.8) in the IBS group (P = 0.6), the mean volume of platelets used in the control group was 0.56 (±SD 0.94) compared to 1.0 (±SD 2.2) in the IBS group (P = 0.39). Laboratory coagulation tests were taken for those patients on arrival in ICU, the mean INR (international normalized ratio) in the control group was 1.42 (±SD 0.19) compared to 1.48 (±SD 0.44) in the IBS group (P = 0.78). One patient in each group had to return to the theatre for haemostasis.
The mean hospital stay for the cell saved survivors was 15.7 days (±SD 9.17; median 13 days) and the mean stay for control group survivors was 14.6 days (±SD 6.8; median 13 days, P = 0.79).
Cost analysis as by the method of Haynes et al.14
Mean costs (£)
If the costs incurred for payment of dedicated technicians to run the CS machine had been included, then the costs obviously would no longer be favourable to the use of the machine; however, many institutions (ours included) have perfusionists on an on-call rota and we believe other staff (such as anaesthetic practitioners) can be trained to use the machine as part of their existing job descriptions without incurring extra cost.
Autologous transfusion was first used in abdominal trauma by Van Schaik in 1927.15 During the Second World War, autotransfusion was utilised by mopping, ladled, or suctioning blood in an open system. Anticoagulation was accomplished with haemodilution or citrate, and cheesecloth or fine gauze was used for filtration. Of a series of 100 cases of trauma to the thorax or abdomen in which autotransfusion was used, 67 were in shock when first seen, and the amount of blood autotransfused ranged from 200 to 2,500 mL, the average amount being 950 mL. Of these 67 patients, 25 had sustained abdominal wounds with intestinal perforations. There was 1 death and 1 reaction to the autotransfusion in this group of patients. The 1 fatality occurred 62 hours after operation. This patient’s injury included 8 perforations of the small intestine, 3 of the descending colon, and 2 of the rectum. He received 2/L of contaminated blood; the transfusion was discontinued when the infusion needle became plugged with faeces. The conclusion, based on the 66 survivors, was that transfusion of grossly contaminated blood was more beneficial than giving no blood at all.16
In 1978, Glover et al. described 183 emergency operations using IBS with 14 patients being transfused with enteric-contaminated blood. Of the 14 patients, 8 survived; all had received antibiotics.17 In the same year, Smith et al. showed that addition of antibiotics to severely hypovolaemic dogs transfused with contaminated blood increased survival from 30% to 90%.18
In 1984, Jurkovich et al. reported on their experiences with 85 trauma patients in whom the use of IBS was attempted. Blood contaminated by enteric contents was reinfused in 6 patients, 4 with colon injuries and 2 with upper gastrointestinal injuries. Cultures of specimens obtained from the cell saver reservoir were positive for Escherichia coli in 2 and α-Haemolytic streptococcus in 1. Five out of the 6 survived, with 1 patient succumbing to sepsis-related multiple organ failure 3 weeks postoperatively. None of the patients; however, had positive blood cultures in the first postoperative week.19 In 1988, Timberlake and McSwain reported the use of cell savers in 11 patients with penetrating torso wounds who received autotransfusion of enteric-contaminated blood. IBS blood comprised 45.5% of the total transfusion requirement for these patients. All patients were treated with broad-spectrum antibiotics. One patient developed necrotising fasciitis of the abdominal wound, but there was no intra-abdominal sepsis and there were no deaths.7
In 1992, the same group published a non-randomised, prospective series of 70 patients with severe abdominal injury. Infection rates were stratified using the PATI. Fifty patients with a PATI score > 20 who received allogeneic blood were compared to 20 patients who received salvaged blood that was potentially infected. The wound infection rates were identical (25% in both groups) and no correlation was found between organisms cultured from the salvaged blood and those causing postoperative bacteraemias, pulmonary or urinary infections.8
In 1992, Horst et al.10 reviewed the use of IBS in a series of 154 trauma patients. Of a total of 66 patients with bowel injury, 58 patients received contaminated blood. The patients who received contaminated IBS developed coagulopathy at lower levels of transfusion compared to patients receiving non-contaminated blood. However, the 117 patients who received less than 10 units of IBS blood had minimal changes in clotting parameters irrespective of the presence of bowel injury. Horst et al. concluded that approximately 3/L of IBS blood can be given without severe coagulopathy, and they recommended a conservative ceiling of just over 2/L of IBS blood for patients with bowel injury.
In their study, complications relating to infection developed both in patients with and without bowel injury; in fact, the infection rate was actually higher in patients without bowel injury.
Successful use of culture-positive cell saver blood has also been demonstrated in other surgical specialities. In a review of autologous transfusions conducted over 2 years in Japan, a sample of autologous blood was taken before reinfusion; in fully one-third of all instances where IBS was used, the blood produced a positive bacterial culture.20 In a prospective study of 401 patients undergoing cardiac surgery, Ezzedine et al.21 cultured samples of the salvaged blood from every case and found an incidence of infection of 12.7%. Locher and Sailer22 used IBS in maxillofacial surgery and calculated that most patients received between 1 and 10 million organisms in the volume of blood retransfused. In 1998, Jeng et al. reported 8 patients where IBS was used during excision of major cutaneous burns. All the reinfused blood was contaminated with bacteria, without adverse clinical consequences.11
This study demonstrates that postoperative sepsis is a powerful predictor of a poor outcome as only a third of patients who had documented episodes of sepsis survived compared to 78% of patients who did not have a septic episode; however, those patients who had IAS blood were no more likely to have sepsis compared to those who received allogeneic blood.
The other major concern using cell-saved blood is the promotion of coagulopathy; we found no excess of coagulopathy in the IBS group; similar volumes of FFP and platelets were transfused and initial laboratory parameters were no different between those patients who arrived in the ICU.
The cost effectiveness of using autologous blood transfusion is hotly debated.5 Some studies have taken into account increases in postoperative infections mediated by immunomodulation with allogeneic but not autologous blood and have found autologous transfusion to be at least cost effective and perhaps cost saving.23,24 Other studies have found that autologous transfusion is not cost effective.25 In this study, as in others,14 IBS has proven to be cost neutral, although we concede that costs incurred to provide technical staff to run the machine could be prohibitive.
Provision of blood in a military or austere setting is beset with logistic difficulty, casualties often requiring fresh blood drawn from their colleagues.26 Supply lines may be extended; for example, of the hundreds of units of blood sent from UK to the Falkland Islands, none were delivered to the Field Hospital on the Falklands until the war was over.27 In military deployments short of war, wastage of blood is a major problem. Between 1995 and 2000, approximately 5,300 units of blood were sent to Bosnia to support American military forces, the usage rate of which was 1.7%.27 IBS has the potential to provide effective, safe blood for transfusion in these settings.
Intraoperative blood salvage has been shown to effectively reduce the demand for bank blood after abdominal trauma by 30% to 45%;4,28 however, presence of contamination has been considered an absolute contraindication to the use of IBS;5 the results of this trial suggest that this is not the case. Consideration should be given to the use of IBS in patients with penetrating abdominal trauma if suitable bank blood is not available. Furthermore, IBS could be considered in those patients with penetrating trauma whose religious beliefs preclude allogeneic transfusion but not autotransfusion.29