Introduction
Annually, an estimated 1.25 million patients undergo cardiac surgery worldwide.(1) In this population, consumption of blood products is reported to be higher than in other surgical groups.(2,3) Cardiac surgery is often accompanied by acute blood loss which affects blood volume, haemoglobin (Hb) and clotting factors. The premise behind blood transfusion is to replace volume, enhance oxygen delivery and improve haemostasis.(4)
Transfusions are increasingly recognised as a risk factor for adverse outcome after cardiac surgery.(5) Several studies have shown an association between blood transfusion, morbidity and mortality in patients undergoing cardiac surgery,(6–12) with transfusion found to be an independent predictor of poor outcome.(13,14)
Considerable variations exist in transfusion practices between different countries and between cardiac surgery centres within countries.(15–19) Red blood cell (RBC) transfusion rates varying from 9% to 100%, fresh frozen plasma (FFP) from 0% to 98% and platelets transfusion from 0% to 51% intraoperatively have been reported.(20) This demonstrates a lack of consensus, and hence, continuous review of these practices is needed.(20) Evidence suggests that transfusion guidelines reduce unjustified transfusions,(15,16,21) cost and make appropriate utilisation of blood products mandatory.(22,23)
Promoting appropriate blood use is one of the major objectives of the World Health Organization (WHO). This can be achieved by monitoring of blood use through serial audits, development of local guidelines and protocols and incorporating haemovigilance programmes in surgical centres.(24) This study thus sought to audit blood transfusion practice in a cardiothoracic unit as an initial step in developing a protocol for appropriate blood use.
Methods
This was a descriptive, retrospective audit conducted at Charlotte Maxeke Johannesburg Academic Hospital, a 1088-bedded quaternary referral centre in Johannesburg. Approximately 160 adult cardiac surgeries on cardiopulmonary bypass (CPB) are performed at this institution annually.
Consecutive records of 122 adult patients who underwent their first elective cardiac surgery on CPB were analysed. Anonymised data were retrieved from records completed by the anaesthetist, perfusionist and intensive care unit doctors. The data collected included demographics, perioperative blood profile, aspects regarding the use of CPB and the blood products used. The information collected was used to create a single database, and the data were collected by one author (CP).
A descriptive statistical analysis was performed using Microsoft Excel™ 2007. Descriptive statistics included the mean, standard deviation (SD), median, interquartile range (IQR), frequencies, numbers and percentages where appropriate. Approval to conduct the study was obtained from the local Human Medical Research Ethics Committee.
Results
A total of 138 patients underwent cardiac surgery on CPB between January and December 2013. Fifteen participants (12.2%) were excluded. Amongst those excluded, 3 had emergency surgery, and 12 had missing data. Records of 122 patients were thus eligible for analysis. A summary of the patient demographics is shown in Table 1.
Demographic and surgical data
| Demographics | Mean | SD |
|---|---|---|
| Age (years) | 46.7 | 16.2 |
| Weight (kg) | 70.9 | 17.9 |
| Height (m) | 1.6 | 0.1 |
| BMI (kg/m2) | 21.4 | 5.3 |
| Gender | N | % |
| Male | 74 | 60.7 |
| Female | 48 | 39.3 |
| Diagnosis | N | % |
| Coronary artery disease | 28 | 22.9 |
| Valve disease | 83 | 68.0 |
| Myxoma | 4 | 3.3 |
| Aneurysm | 7 | 5.7 |
| Surgical procedures | N | % |
| Single valve repair/replacement | 36 | 29.5 |
| Double valve repair/replacement | 37 | 30.3 |
| Triple valve repair/replacement | 7 | 5.7 |
| Isolated CABG | 28 | 22.9 |
| Single valve repair and CABG | 3 | 2.5 |
| Myxoma removal | 4 | 3.3 |
| Aneurysm repair | 7 | 5.7 |
BMI = body mass index, CABG = coronary artery bypass graft.
Hb levels and platelet counts were assessed perioperatively (Table 2).
Perioperative haemoglobin and platelet counts
| Blood results | Mean (SD) | Range |
|---|---|---|
| Haemoglobin (g/dL) | ||
| Preoperative | 12.8 (2.3) | 7.7–18.8 |
| Intraoperative: First on CPB | 8.9 (1.6) | 5.4–16 |
| Last on CPB | 9.9 (1.3) | 6.8–13 |
| Postoperative: Intensive care unit | 11.1 (1.6) | 7–13.5 |
| Platelets 10 9 /L | ||
| Preoperative | 274.4 (121.9) | 34–838 |
| Postoperative | 172.9 (71.3) | 42–308 |
CPB = cardiopulmonary bypass.
In the current study, anaemia, defined as Hb levels <12.0 g/dl in women and <13.0 g/dl in men according to the WHO,(24) was found to be present in 48 (40%) patients and thrombocytopaenia in 11 (9%) patients preoperatively. The mean (SD) preoperative platelet count was 274.4 × 109/l (121.9). Postoperatively, the mean (SD) platelet count was 172.9 × 109/l (71.3). A small proportion (8%) of patients had platelet counts below 150 × 109/l postoperatively. Overall, 57 (46.7%) patients received preoperative anticoagulation therapy. Of these, 36 (63.1%) were on Coumadin while 21 (36.9%) were on aspirin and enoxaparin.
The CPB machine was primed with a crystalloid in all patients. The median (IQR) priming volume was 1500 (1000–2000) ml. The anticoagulant used on CPB in all the patients was heparin, and the mean (SD) dose given was 90 (0.91) mg.The overall median (IQR) time on CPB was 117 (95–145) min. In total, 97 (79.5%) received cell-salvaged blood, and the median (IQR) volume of salvaged blood was 535 (250–754) ml. The time on CPB and volume of blood salvaged for each surgical group are presented in Table 3.
Time on cardiopulmonary bypass and blood salvaged.
| Type of Surgery | Time on CPB (min) median (IQR) | Cell saved blood (ml) median (IQR) |
|---|---|---|
| Single valve repair/replacement | 113 (88–135) | 682 (108–875) |
| Double valve repair/replacement | 117 (96–139) | 600 (327–900) |
| Three valve repair/replacement | 145 (135–170) | 306 (250–650) |
| Isolated CABG | 117 (97–150) | 500 (200–678) |
| Single valve plus CABG | 115 (102–202) | 600 (570–700) |
| Myxoma removal | 75 (62–98) | 426 (184–712) |
| Aneurysm repair | 130 (81–240) | 500 (0–700) |
CPB = cardiopulmonary bypass, CABG = coronary artery bypass graft.
Overall, 110 (90.2%) patients were transfused with donor blood products. Among those transfused, 85 (77.3%) received RBCs, 103 (93.6%) FFP and 35 (31.8%) received platelet transfusion. The mean (SD) number of units transfused per patient was 3.14 (2.2) RBC, 1.8 (1.5) FFP and 0.3 (0.4) platelet units.
A breakdown of product units used per patient according to the type of surgery performed is shown in Table 4.
The number of product units per patient per surgical group.
| Type of surgery | Received blood products n(%) | RBC mean (SD) | FFP mean (SD) | PLT median (IQR) |
|---|---|---|---|---|
| Single valve repair/replacement | 33 (91.7) | 2.2 (2) | 1.9 (0.5) | 0 (0-1) |
| Double valve repair/replacement | 33 (89.2) | 1.6 (1.9) | 1.6 (1.2) | 0 (0-2) |
| Triple valve repair/replacement | 7 (100) | 2.6 (0.9) | 1.6 (0.9) | 0 (0-0) |
| Isolated CABG | 23 (82) | 1 (0-3)# | 1.6 (1.3) | 0 (0-0) |
| Single valve plus CABG | 3 (100) | 3 (2.6) | 3.3 (2.3) | 0 (0-2) |
| Myxoma removal | 4 (100) | 4 (3.2) | 2.6 (1.6) | 0.5 (0-1) |
| Aneurysm repair | 7 (100) | 2 (0-4)# | 2.6 (2.2) | 0 (0-1) |
RBC = red blood cell, FFP = fresh frozen plasma, PLT = platelet, CABG = coronary artery bypass graft.
Discussion
Transfusions are increasingly recognised as a risk factor for adverse outcome after cardiac surgery.(6,8,9) In the current study, just over 90% of the patients were transfused with donor blood products, with the majority (77.3%) receiving RBC transfusions. Snyder-Ramos et al. (20) studied intraoperative transfusion rates in 16 countries. The average rate of transfusion of all blood products was much lower in these 16 countries with rates of 41.5% for RBC, 10.9% for FFP and 9% for platelet transfusions. However, rates of transfusion varied significantly amongst the 16 countries. The intraoperative RBC transfusion rates varied from 9% to 100%, FFP from 0% to 98% and platelet transfusion from 0% to 51%.(20) Similarly, a multicentre audit of 25 Australian hospitals showed transfusion rates ranging between 22% and 67% for RBCs, 11% and 48% for FFP and 11% and 39% for platelet units, respectively.(25) A national audit of 25 cardiac centres in the United Kingdom reported variations of transfusion rates of 20%–60% for RBC and 5%–45% for FFP.(17)
Some countries have shown a decline in blood transfusions over time. An audit of RBC transfusion rates for coronary artery bypass surgery in Denmark reported a progressive reduction in transfusion rates from 62.2% (2004), 55.8% (2008) to 46.9% in 2010.(26) Another study in the United States in 2011 reported an overall transfusion rate of 43%.(27) In 2015, Geissler et al. (19) conducted a large single-centre retrospective audit of their cardiac surgical centre in Germany and reported transfusion rates of 50% for RBC, 25% for FFP and 20% for platelet units.
Analysis of blood use by surgical procedure in the current study showed that the single valve repair patients received the highest number of RBC (30.9%) and FFP (31.1%) transfusions compared to other surgical groups. The group that received the least number of RBC and FFP (3.5% and 4.4%, respectively) transfusions was the combined single valve plus CABG surgery. On the other hand, Scott et al. (28) reported a higher transfusion rate (69.8%) in the CABG compared to lower transfusion rates (33.9%) found in the valve surgery group.
The vast differences in blood product use internationally have been attributed to different patient populations, differences in procedure-related factors, traditions and norms associated with transfusions, the use of variable protocols and lack of adherence to international recommendations.(20)
Independent preoperative clinical variables that predict the likelihood of exposure to blood transfusion in patients undergoing cardiac surgery have been identified. These variables include preoperative Hb, weight, female gender, emergency surgery and combined procedures.(28,29) The mean age of 46.7 (16.2) years of the population in the current study was relatively young and predominantly male. The mean preoperative Hb was 12.8 (2.6) g/dl, and the body mass index (BMI) was low at 21.4 (5.6) kg/m2. In comparison, the patient profile in a comparative study had older patients at 64.1 (9.8) years, who were relatively overweight with a BMI of 27.6 (4.5) kg/m2, yet they had lower transfusion rates.(20) Some studies also reported lower transfusion rates than that found in the current study despite including emergency surgeries as part of their inclusion criteria.(20,27,28)
Bleeding in cardiac surgery patients is multifactorial. Transfusion of FFP and platelets is used to manage bleeding. The FFP and platelet transfusion rates were found to be 93.6% and 31.8%, respectively, in the current study. Glasgow et al.(30) found FFP transfusion rates to be 38.3% in their study. In their study, the use of FFP and platelet was guided by point-of-care testing.(27) Standard laboratory tests for assessing coagulopathy are impractical to carry out intraoperatively; however, point-of-care devices, such as thromboelastogram, are able to assess global coagulopathy and yield results within minutes. In the setting of the current study, point-of-care testing for transfusion was not used, and FFPs and platelets were transfused without prior testing to guide therapy, suggesting that transfusions were given empirically.
Given the low-risk nature of the population in the current study, and that emergency surgery was excluded, the risk for transfusion was anticipated to be lower; however, it was found to be higher than in many previously reported studies. Management of Hb was consistent with practice in other studies, but this occurred at the expense of increased rates of transfusion. There are multiple explanations for the increased rates of transfusion, but one of the major reasons is poor blood-conservation strategies during surgery. In this study, there was use of high volumes of cell-salvaged blood. The use of high volumes of cell-salvaged blood meant that high volumes of blood were collected from the surgical field and directed away from the CPB pump reservoir, potentially leading to less volume in the pump prime. Less volume in the pump would need an increase in on-pump prime with clear fluids, leading to a drop in Hb. To circumvent the drop in Hb on pump, the perfusionist would need to use autologous blood at this point to maintain pump prime and Hb. Furthermore, blood processed though the cell saver loses platelets, plasma volume and plasma proteins including clotting factors. Deficiency of these in turn would lead to an increase in transfusion rate of these components. Furthermore, there was no stipulated single transfusion trigger due to the absence of a protocol, blood-conservation strategy and point-of-care testing. In the absence of point-of-care testing, FFP and platelet transfusions were empirical and sometimes may have to be unjustified. Furthermore, the decision to give blood is often dependant on the team, individual surgeon or perfusionist. All these factors probably contributed to the transfusion rates observed.
The mean Hb of 8.6 g/dl in this study was higher than that used as an endpoint in the restrictive group of the Transfusion Requirements in Cardiac Surgery (TRICS) III trial.(31) In that trial, in moderate-to-high-risk patients, the restrictive arm of the study used an Hb of 7.5 g/dl as a cut-off for transfusion of red cells intraoperatively during cardiac surgery. The strategy was found to be non-inferior to a liberal strategy using an Hb of 9.5 g/dl as a cut-off. Thus, one of the steps to designing a locally adapted transfusion protocol is to use a lower Hb (7.5 g/dl) to initiate red cell transfusion.
Clinical complications posed by transfusion of blood have led to the development of blood-conservation strategies in cardiac surgery.(26) Studies have summarised conservation strategies as preoperative autologous blood donation, use of erythropoietin, intraoperative use of antifibrinolytic agents, cell salvage and acute normovolaemic haemodilution.(2,29) The Society of Cardiothoracic Surgery and The Society of Cardiovascular Anesthesiologists in the United States have developed guidelines for blood transfusion, with specific consideration for the cardiac patient.(2) The use of these guidelines has been credited with significant improvements in practice and cost reductions.(22,26) The routine use of the cell salvage technique and its cost-effectiveness in low-risk patients has been challenged.(28)
The cell salvage and reinfusion technique was used in 94 (77.0%) patients in the current study. The median volume of blood reinfused was 535 ml (250–764). In comparison, an overall autologous transfusion frequency of only 9% was shown in a 2008 multicentre Snyder-Ramos et al. study.(20) In one country in that study, no patient received autologous blood. In Scott et al., (28) cell salvage rates of 12.3% were reported and attributed these rates to selective use of the cell salvage method, which was reserved only for high-risk patients.(28) Reports regarding the use of cell salvage in the literature are conflicting, but the consensus is that in the correct setting, its use is associated with a decreased rate of transfusion.(26)
Preoperative anaemia was found to be present in 48 (40%) of the patients in the current study. Scott et al. (28) reported preoperative anaemia to be present in 24.2% of their population, and their blood transfusion rates were lower. In the current study, the mean Hb in the intensive care unit was found to be 9.9 g/dl, and 82% had a postoperative Hb of >10 g/dl. Oxygen delivery and metabolic demands vary with age, BMI, gender, comorbidities and the presence of active bleeding. Thus, a single universal number cannot describe anaemia.(15)
In the current study, thrombocytopaenia was found in 11 (9%) patients preoperatively; however, none of the thrombocytopaenic patients were below the bleeding threshold. A total of 35 (31.8%) patients received platelet transfusions intraoperatively despite the lack of intraoperative assessment of platelet numbers or function.
Limitations
The current study should be interpreted with the following limitations in mind. First, this was a retrospective analysis. Also, this study did not assess possible wastage of blood products after it was issued from the blood bank. A further limitation is that this study did not study the relationship between the current transfusion practices and clinical outcomes as it has been shown that there is increased risk of morbidity and mortality with blood transfusions.
Conclusion
Despite the patient population being of low-risk and the fact that emergency surgery patients were excluded, a high rate of homologous blood product transfusion was found in the study. Lack of institution-specific guidelines, use of cell salvage leading to high volumes of cell saved blood being re-transfused but directed away from the CBP reservoir, a higher Hb threshold before initiating transfusion and lack of use of point-of-care devices to assess coagulopathy may have contributed to the high transfusion rate.
Recommendation
First, an institution-specific blood transfusion guideline needs to be established by all role players involved in the care of patients undergoing cardiac surgery. This guideline should incorporate preoperative identification of patients with high-risk profiles associated with increased perioperative risk of transfusion with blood products, interventions that are likely to decrease the use of blood products in such patients and blood transfusion algorithms together with point-of-care testing.
Based on the current study, acceptance of a lower on-pump Hb (7.5 g/dl), shown not to be inferior to a higher Hb (9.5 g/dl) in the TRICS III trial, should perhaps be incorporated into the local guidelines. Employing the use of a point-of-care device, changing CPB priming techniques and using the cell salvage reinfusion technique only in selected high-risk patients in our institution may lead to a change in the current blood transfusion practice. Further studies are suggested to assess the effect of these guidelines on transfusion rates as well as its link with clinical outcomes.