Autotransfusion with cell saver for off-pump coronary artery bypass surgery: a randomized trial

Abstract

Objectives. Off-pump coronary surgery reduces transfusions, however, many patients still receive blood. This trial aims to clarify the effect of using a cell saver intraoperatively. Design. In 60 patients shed blood was collected in the cell saver reservoir intraoperatively; randomization and processing or discharge were performed immediately after surgery. Primary outcome measures: proportion of patients receiving allogeneic blood, and average number of units per patient. Secondary outcome measures: blood loss, hemoglobin levels, complications, and costs. Results. Cell saver group versus control group; received transfusions: 17/30 vs. 14/29 (p = 0.28), allogeneic units: median 1 (interquartile range 0 – 2) vs. 2 (IQR 0 – 7) (p = 0.06), intraoperative net blood loss: median 300 ml (IQR 193 – 403) vs. 610 ml (IQR 450 – 928) (p < 0.001). Control group patients had more complications leading to transfusion. Hemoglobin levels and costs were comparable between groups. Conclusions. Use of cell saver reduced intraoperative net blood loss and seemed to reduce transfusions by 1 unit per patient, however, this was probably attributable to more complications leading to transfusion in the control group. In the future larger trials are necessary.

In developed countries allogeneic blood transfusions carry low, but well known risks of primarily infectious (viral) and immunological complications. In cardiac surgery additional concerns have been raised about especially bacterial infections, and impaired short and long-term survival following transfusion 1–3. Not only risks but also substantial costs and increasing lack of donors have prompted development of alternatives to allogeneic blood transfusions.

Antifibrinolytic agents and especially aprotinin have proved efficient for reducing transfusions in cardiac surgery 4. Transfusion of washed autologous blood using a cell saver also seems to reduce allogeneic transfusions in cardiac surgery with cardiopulmonary bypass (CPB), while the evidence for efficacy of autotransfusion of unwashed blood is scarce 5. Likewise evidence on safety and efficacy of low hemoglobin thresholds for transfusion is limited 6. Finally, cardiac surgery without CPB: off-pump coronary artery bypass grafting (OPCAB) has proved to reduce allogeneic blood requirements 7. But in our experience blood loss is still substantial during OPCAB surgery, approximately 500 – 1000 ml. A retrospective survey at our department showed that this led to transfusion in about half of patients or 2 – 3 units of allogeneic blood components per patient, even with administration of tranexamic acid and postoperative autotransfusion of drainage blood. The role of cell saver for OPCAB is undefined, but with intraoperative use of cell saver in an uncontrolled prospective design, we have noticed a decrease of approximately 1 unit of transfused blood component per patient. It is this apparently positive effect of cell saver in OPCAB surgery we wish to test in a randomized trial design.

Material and methods

Design

Randomized clinical trial including 60 patients. Method of randomization: 60 sealed and opaque envelopes numbered in sequence randomly allocating 1:1, 30 patients to the study (cell saver) group and 30 to the control group.

Primary outcome measures

The proportion of patients receiving transfusion with allogeneic blood components (units of red blood cells (RBC), fresh frozen plasma and pooled platelets), and average number of units transfused per patient during the admission.

Secondary outcome measures

Intraoperative and postoperative bleeding, changes in hemoglobin and hematocrit, and registration of complications and costs.

Patient inclusion

Patients admitted for elective or sub-acute* coronary bypass surgery without heart-lung machine were included (*Operated during the same admission after coronary angiography). Acute patients from the catheterization laboratory, patients presently in anticoagulation therapy with coumarine derivatives, with thrombocytopenia (< 150×10⁹), prior cardiac surgery or Jehovah's Witnesses were excluded.

There were no limitations regarding age, ejection fraction (EF) or number of vessel disease. If cardiopulmonary bypass became necessary during the operation, the patient was excluded. All patients gave oral and written informed consent according to the regulations of the science ethics committee.

Surgery, anesthesia and cell saver

Anesthesia and surgery followed current guidelines of the department, this included partial heparinization (100 International Units/kg) and partial reversion with protamine sulphate (0.5 mg/IU heparin). Surgery was performed through median sternotomy and cardiac stabilization with Medtronic Starfish and Octopus. Generally, the left internal mammarian artery was anastomosed to the left anterior descending artery and saphenous vein or radial artery was used for free grafts.

During surgery the continuously heparinized suction (300 IU of heparin in 1000 ml saline) and reservoir belonging to the cell saver (Medtronic, AutoLog) were used for all patients in both groups. Immediately after surgery the patients were allocated to the study group (cell saver) or to the control group by opening the envelope numbered according to the patient's study number assigned at inclusion. The suctioned blood from patients in the cell saver group was processed and autotransfused before the patient was transferred to the intensive care unit (ICU), while the control group had their suctioned blood discharged. In this way, the surgical and anesthetic team were blinded during the operation, but not after. However, the ICU and ward personnel were not informed about which procedure had been performed. The cell saver reservoir with a 40-micron filter was used at the ICU for mediastinal drain blood collection and for postoperative autotransfusion in all patients, if possible (adequate volume and quality). A maximum of 12 hours of postoperative unwashed autotransfusion from the drains is routine.

Both groups received allogeneic blood transfusion when indicated and when drain blood volumes in the reservoir were inadequate for autotransfusion. Indication for RBC transfusion was the usual guidelines of the department: hemoglobin below 6.0 mmol/l and/or hematocrit below 30%. Plasma, platelets and tranexamic acid were given according to clinical judgment. Postoperatively all patients received for thrombosis prophylaxis: enoxaparin 4000 IU per day subcutaneously for four days and aspirin 75 mg daily, unless ongoing drain bleeding prohibited this.

Statistical analysis

Data were handled in Microsoft Access and analysed with SPSS version 13. Averages were compared with the non-parametric Mann-Whitney U test and presented as medians and interquartile range (IQR-25th to 75th percentile), and range where appropriate. Proportions were compared using Fisher's exact test and presented as counts and percentages. P-values are expressed as two-sided and values below 0.05 are considered significant.

Ethical considerations

The study was carried out in accordance with the Helsinki Declaration of 1975 as revised in 1983, and approved by the local science ethics committee on 24 July 2003, registration number: (KF) 01-110/03.

Results

Sixty patients were included (September 2003 to October 2004) and 30 allocated to the cell saver group and 30 to the control group. Overall it was a population of old age and high risk patients: median age 77 years (IQR 73 – 79, range 41 – 86) and median EuroScore of 6 (IQR 5 – 7, range 0 – 12). Baseline characteristics showed no differences. Preoperative medication on admission was likewise comparable in the two groups, patients were considered not taking aspirin if there had been a seven day preoperative pause (Table I).

Table I.  Baseline characteristics.

	Cell saver (n = 30)	Control (n = 30)	p value
Median (interquartile range)
Age	77 (74–79)	76 (70–79)
EuroSCORE	6 (5–7)	6 (5–7)
Ejection fraction, %	58 (50–60)	50 (40–60)	0.08
Hemoglobin, mmole/l	7.9 (7.4–8.7)	8.2 (7.4–8.9)	0.43
Hematocrit, %	39 (36–42)	41 (38–44)	0.21
Thrombocytes, 10^9/l	236 (186–297)	259 (200–330)	0.39
Factor II + VII + X, INR	1.0 (1.0–1.2)	1.0 (1.0–1.1)
Number (%)
Female sex	11 (37)	14 (47)
3-vessel disease	23 (77)	19 (63)
Previous stroke	6 (20)	3 (10)
Chronic obstructive pulmonary disease	5 (17)	4 (13)
Diabetes	3 (10)	5 (17)
Renal failure (creatinine >200 µmole/l)	3 (10)	1 (3)
Aspirin	15 (50)	12 (40)
Low molecular weight heparin	9 (30)	11 (37)
Clopidogrel	1 (3)	0

Intraoperatively one patient in the control group was excluded due to conversion to conventional coronary surgery using CPB because of insufficient EF for OPCAB procedure. Due to logistic or technical problems the cell saver was not used for five patients in the cell saver group. According to intention to treat principles, they were kept in the study analysis. There were no differences in intraoperative events except a trend towards lower hemoglobin and hematocrit in the cell saver group (Table II).

Table II.  Intraoperative events.

	Cell saver (n = 30)	Control (n = 30)	p value
Median (interquartile range)
Duration of operation, min	165 (135–186)	150 (135–188)	0.39
Heparin, ×100 IU	98 (75–100)	87 (75–100)	0.46
Protamine, mg	50 (50–50)	50 (30–50)	0.24
Minimum patient temperature, °Celsius	35.1 (34.5–35.5)	35.0 (34.6–35.5)
Total bleeding, ml	655 (508–818)	610 (450–928)	0.56
Net blood loss, ml	300 (193–403)	610 (450–928)	<0.001
Cell saver red blood cell suspension, ml (n = 25)	126 (120–150)	–
Red blood cell suspension hematocrit, % (n = 15)	49 (27–67)	–
Hemoglobin, mmole/l (lowest)	6.0 (5.7–7.1)	6.5 (5.9–7.3)	0.09
Patient hematocrit, % (lowest)	29 (29–34)	32 (29–37)	0.06
Activated coagulation time, sec (highest)	259 (207–283)	257 (208–292)	0.95
Number (%)
Non-elective operation	19 (63)	15 (52)	0.44
LIMA/RIMA	29 (97) / 2 (7)	27 (93) / 1(3)
Saphenous vein grafts: 1/2/3	7 / 14 / 2	10 / 14 / 0
Radial artery grafts: 1/2/3	4 / 2 / 1	2 / 1 / 0	0.18
Atrial arrythmia	4 (13)	4(14)
Ventricular arrythmia	2 (7)	0
Inotropic bolus	26 (87)	23 (79)
Inotropic infusion	4 (13)	9 (31)	0.13
Tranexamic acid	13 (43)	15 (52)	0.61

Use of the cell saver led to a 310 ml lower median net blood loss in the cell saver group (p < 0.001), since cell saver group patients received median 126 ml of processed autologous RBC suspension from the cell saver (Table II).

Postoperatively two patients in the control group, 80 and 79 years old with EuroSCORE 7, died day 5 and 31 of irreversible cardiac arrest, both after complicated admissions including multiple allogeneic transfusions and dialysis. The patients in the control group had more postoperative complications: six versus one stayed in the intensive care unit for more than 24 hours postoperatively (p = 0.05), and six versus none had low cardiac output syndrome* postoperatively (p = 0.01) (*Infusion of epinephrine/ norepinephrine at a rate above 0.05 µg/kg/hour in more than 1 hour after arrival to the intensive care unit, to maintain a systolic bloodpressure above 90 mmHg). There were no differences between groups in drain bleeding and autotransfusion, or hemoglobin concentration and hematocrit throughout the admission (Table III).

Table III.  Postoperative events.

	Cell saver (n = 30)	Control (n = 29)	p value
Median (interquartile range)
Days of admission	7 (6–8)	7 (6–9)
Total drain bleeding, intensive care, ml	770 (593–964)	788 (600–1205)	0.32
Autotransfusion, intensive care, ml (n = 14 vs. n = 13)	463 (325–555)	700 (325–850)	0.37
Peak creatin kinase MB, g/l	12 (7–18)	14 (8–24)	0.35
Hemoglobin, intensive care (lowest), mmole/l	5.9 (5.3–6.6)	5.8 (5.2–6.7)	0.97
Hematocrit, intensive care (lowest), %	29 (27–33)	29 (25–33)	0.69
ACT, intensive care (highest), sec	106 (93–127)	101 (93–110)	0.39
Hemoglobin, ward (lowest), mmole/l	6.4 (5.9–6.8)	6.6 (5.8–7.1)	0.58
Hemoglobin, at discharge, %	7.1 (6.5–7.4)	7.2 (6.5–8.1)	0.25
Number (%)
Deaths	0	2 (7)	0.24
Stroke	0	1 (3)
Myocardial infarction	0	1 (3)
Reoperation for bleeding	1 (3)	3 (10)	0.35
Pneumonia	2 (7)	3 (10)
Gastrointestinal bleeding	0	3 (10)	0.11
Deep sternal wound infection	0	1 (3)
Leg wound infection	0	1 (3)
Dialysis	1(3)	2 (7)
Creatinine > 200 mole/l	4 (13)	7 (24)	0.33
Ventilator > 24 hours	0	3 (10)	0.11
Intensive care unit > 24 hours	1 (3)	6 (21)	0.05
Low cardiac output syndrome*	0	6 (21)	0.01
Atrial arrythmia	14 (47)	20 (69)	0.12
Ventricular arrythmia	0	3 (10)	0.11
Inotropic infusion	6 (20)	9 (31)	0.38

*Infusion of epinephrine/norepinephrine at a rate above 0.05 µg/kg/hour in more than 1 hour after arrival to the intensive care unit, to maintain a systolic bloodpressure above 90 mmHg.

Primary outcome measures

17/30 patients in the cell saver group and 21/29 in the control group received allogeneic blood components (RBC, plasma and platelets) (p = 0.28). The median total amount of allogeneic transfusions per patient seemed lower in the cell saver group: 1 unit (IQR 0 – 2) compared to the control group: 2 units (IQR 0 – 7) (p = 0.06). Median RBC units (cell saver versus control): 1 unit (IQR 0 – 2) versus 2 units (IQR 0 – 5) (p = 0.07). The average cost for patients in the two groups was calculated on the basis of the mean amount of units per patient in relatively uncomplicated cases, excluding reoperated and dead patients. Without doing this, the cost calculation would erroneously be in overwhelming favour of the cell saver group. The corrected calculation showed comparable costs in both groups (Table IV).

Table IV.  Transfusions and costs.

	Cell saver	Control	P value
	(N = 30)	(N = 29)
Median (interquartile range; range)
Red blood cells, intraoperative	0 (0–0; 0–2)	0 (0–1; 0–3)	0.13
Red blood cells, ICU	1 (0–2; 0–13)	1 (0–3; 0–27)	0.30
Fresh frozen plasma, ICU	0 (0–0; 0–4)	0 (0–0; 0–22)	0.40
Pooled platelets, ICU	0 (0–0; 0–1)	0 (0–0; 0–1)
Red blood cells, ward	0 (0–0; 0–2)	0 (0–1; 0–6)	0.22
Fresh frozen plasma, ward	0	0 (0–0; 0–1)	0.31
Total, allogeneic units	1 (0–2; 0–17)	2 (0–7; 0–49)	0.06
Total, red blood cell units	1 (0–2; 0–13)	2 (0–5; 0–27)	0.07
Received transfusion–number (%)	17 (57)	21 (72)	0.28
Without deaths and reoperated patients	(N = 29)	(N = 24)
Total, allogeneic units	1 (0–2; 0–6)	2 (0–3; 0–8)	0.27
Total, red blood cell units	1 (0–2; 0–5)	2 (0–3; 0–7)	0.28
Cost per patient in Danish Kroner (DKR)	(N = 29)	(N = 24)
DKR per unit×mean number of units
Red blood cells	1012×1.28 = 1295	1012×1.92 = 1943
Fresh frozen plasma	224×0.17 = 38	224×0.21 = 47
Pooled platelets	1372×0.03 = 41	1372×0.04 = 55
Cell saver wash kit	525×1 = 525	0
Cell saver suction and reservoir	855×1 = 855	855×1 = 855
Total	2754	2900

Discussion

Not surprisingly our study documents that the intraoperative net blood loss was reduced, since blood collected in the cell saver was processed and reinfused for cell saver group patients and discharged in the control group. Overall, allogeneic and RBC transfusions were comparably reduced by one unit in the cell saver group versus the control group. This could partly be caused by more complications in the control group leading to more transfusions. Statistical analysis (Intention to treat) shows an almost significant trend towards reducing allogeneic transfusional requirements, however, an analysis without the reoperated and dead patients shows no statistical differences (Table IV). Per protocol analysis* gives the same result, no difference (*Only patients treated in accordance with randomization).

Conversely, an effect of the cell saver in reducing transfusions cannot be ruled out. Post hoc sample size calculation based on the present data shows that our study was underpowered to detect a difference in the primary outcome measures: proportion of patients receiving transfusions, and number of allogeneic units transfused per patient. Approximately 450 patients, 225 in each group, would be necessary to detect a 15% reduction of patients receiving transfusion. Alternatively 180 patients, 90 in each group, would be necessary to detect a reduction of 1 unit transfused per patient. These calculations have a significance level of 0.05 and a power of 0.90.

At the time of initiation of this trial there were no reports on this subject. Murphy et al. have recently reported results of a comparable trial including 61 patients, in which no firm conclusions regarding transfusion could be made as well 8.

In conclusion, based on these results, future properly dimensioned trials are needed to fully document the effect and safety of cell saver used for off-pump coronary surgery. Since the use of cell saver is cost neutral and at best reduces blood transfusions, we find there is no reason not to use it, if the equipment is already available.

Sune Damgaard is funded by a research grant from H:S-Copenhagen Hospital Corporation.