Introduction
A significant number of patients in the intensive care unit (ICU) receive packed red blood cell (PRBC) transfusions [
1]. Anaemia which affects up to 90% of ICU patients by Day 3 is multifactorial [
1]. One such cause is blood loss, up to 17% of which is contributed by repeated blood drawing for diagnostic tests [
2,
3]. Blood samples may be drawn up to 24 times in a day, resulting in an average blood loss of 41 ml on Day 1 [
4]. There is a positive correlation between organ dysfunction and the number of blood draws [
2,
3,
5]. The presence of indwelling central venous or arterial catheters makes blood sampling easier but contributes to iatrogenic anaemia as the first few millilitres of infusate-blood mixture obtained while collecting blood from such catheters are discarded [
6‐
8]. In two large trials, 37 to 44% of patients in ICU received PRBC transfusions [
1,
5] often at high transfusion thresholds, despite evidence to support a restrictive transfusion practice to keep haemoglobin (Hb) levels in the range of 7 to 9 g/dL [
9]. Importantly, PRBC transfusions are associated with adverse effects, including allergic, anaphylactic and haemolytic transfusion reactions, transfusion-related acute lung injury (TRALI), transfusion-associated circulatory overload (TACO), acute respiratory distress syndrome (ARDS), infections, and ventilator-associated pneumonia, all of which lead to significant morbidity and mortality [
10‐
14].
Reduction of the discarded blood volume is possible using a three-way connection [
15] or a dedicated blood conservation system [
16]. While data exist to show that such devices may reduce the degree of blood loss [
17,
18] resulting in higher Hb levels [
19], no previous study has demonstrated any significant effect of these devices on the amount of blood transfusion. This apparent paradox may be related to the inadequate sample sizes or study design issues including the lack of standardised thresholds for transfusions [
20]. The primary objective of the present study is therefore to investigate if the use of a blood conservation device in the presence of a standardised restrictive transfusion practice can reduce the number of units of PRBC transfused. The secondary objective is to investigate if the use of the device is associated with a smaller decrease in Hb levels from ICU admission to discharge.
Materials and methods
Study design
This was a before-and-after study conducted in the 12-bed medical ICU of our university hospital. The before-study period included patients from January to June 2008 (control group). The blood conservation device was introduced to the active group at the start of the after period from July 2008 to March 2009 (active group).
Patients
We included all patients admitted to the ICU who were 1) 18 years and above, 2) expected to stay more than 24 hours and 3) had an indwelling intra-arterial catheter inserted. We excluded patients who 1) were expected to stay less than 24 hours and 2) had active gastrointestinal or other bleeding as the primary cause of ICU admission. Patients were followed up till hospital discharge, death or up to 28 days of ICU stay, whichever was later.
Device
We used the Venous Arterial blood Management Protection (VAMP) system (Edwards Lifesciences, Irvine, CA, USA) for the active group. This device has been described previously [
16]. Briefly, it is attached to the existing arterial catheter. While drawing the samples the flexures of the device are firmly squeezed and a blood volume is slowly drawn into the reservoir over three to five seconds. The shut-off valve just proximal to the reservoir towards the patient's end is then closed. The sample site is cleaned and a syringe with a custom-made cannula (Edwards Lifesciences) is attached. A vacuum tube is attached to the syringe and the required blood sample(s) is drawn. Following the collection of the sample, the syringe with the cannula is removed and the shut-off valve is opened. The device's plunger is then pushed down smoothly and evenly over three to five seconds, until the flexures lock in place in the fully closed position and all fluids have been reinfused into the arterial line. A single device was used for an individual arterial catheter throughout the patient's stay and removed or changed with the arterial catheter.
Transfusion practice
We employed a restrictive transfusion practice in both the before and after periods of the study [
9]. Clinicians were strongly discouraged against any routine transfusion of PRBCs when the Hb level was above 7.5 g/dL, unless there was a physiological need for transfusions (including transfusion as part of resuscitation, preoperatively, or in patients with coronary artery disease). Ultimately, however, the decision to transfuse was left to the discretion of the clinicians.
Outcomes
The primary outcome was the number of units of PRBC transfused per patient per day of ICU stay. The secondary outcome was the difference between the Hb levels at ICU admission and discharge.
Data collection
We recorded the following data prospectively: patient demographics, Acute Physiology and Chronic Health Evaluation (APACHE) II score, Hb levels at ICU admission and discharge or death and just before any PRBC transfusion, number of units of PRBC transfused, need for any renal replacement therapy (RRT), ICU length of stay (LOS), and mortality. For the patients who died in ICU, the last Hb before death was recorded.
Sample size
On a ratio of one control to two active patients, with 80% power and a two-sided test of 5%, 80 controls to 160 patients will provide a statistically significant result for a difference of at least 0.05 unit PRBC/patient/day with a standard deviation of 0.15 units PRBC/patient/day.
Statistical analysis
We expressed variables as means ± standard deviations and numbers (percentages), and made comparisons using Student's t-test and the chi-square test where appropriate. To elucidate the independent predictors of transfusion requirement, the following variables were entered into a linear regression model: age, gender, Hb on admission and just before transfusions, LOS, severity of illness, RRT (duration in hours), and use of the blood conservation device. The same variables were entered into a separate logistic regression model to ascertain the independent predictors of ICU and hospital mortality. We used the statistical software SPSS version 17.0 (SPSS Inc., Chicago, IL, USA).
The study was approved by our Institutional Review Board and Ethics committee. Informed consent was obtained in the active group. Requirement of consent was waived for the control group.
Discussion
In the present study, patients using a blood conservation device had a 48% reduction in PRBC transfusion requirements. This was not observed in previous studies using similar devices. The device was also associated with a smaller decrease in Hb levels between ICU admission and discharge.
Use of blood conservation devices has been studied previously. Three-way stopcock and syringes can be used to preserve the discarded blood-infusate [
4]. Silver MJ et al [
19] showed that the blood samples obtained with the blood-conserving arterial line were free of haemodilution or heparin contamination. In a small randomised control trial (RCT), Peruzzi WT et al [
19] showed that the conservation group had better preservation of Hb with less volume of blood being discarded. However, the decrease in the transfusion requirements was not significant. Such devices were also found to be free of microbial contaminations [
21].
Despite their potential benefits, blood conservation devices are rarely used. In a survey of members of the Society of Critical Care Medicine, most agreed that such devices could be very useful in preventing anaemia [
4]. Another survey found that such devices were used in only 18.4% of adult ICUs in England and Wales [
22]. One reason for such a paradox is the lack of convincing data on the effect of these devices on transfusion requirements. Encouragingly, findings of the present study strongly suggest that such devices do indeed reduce PRBC transfusion.
Determination of a transfusion threshold or trigger in the ICU has been challenging. Due to the adoption of a restrictive transfusion practice [
9] in our ICU, only 27.6% of our patient cohort received PRBC transfusions, which is lower than in previous studies [
1,
5]. This is reflected in the similar Hb levels at transfusion in both the control and active (7.1 ± 0.85 vs 7.25 ± 1.1 g/dL) groups. It is likely that concurrent application of the restrictive transfusion practice where transfusion triggers are not individualised but guided, allowed demonstration of the effect of the blood conservation device on transfusion requirements. This notwithstanding, 23.8% and 29.7% patients in the control and active group respectively did receive transfusions above the suggested threshold (Table
2). In addition, a relatively smaller number of patients in the control group (control 17/80, 21.3% vs active 52/170, 30.6%) received a larger number of PRBC transfusions (control 62 units vs active 129 units of PRBC, table
2). This suggests that multiple transfusions of the same patients occurred in the control group.
In our study, the control group had a greater loss of Hb; this finding is consistent with those of previous studies [
15,
18,
19]. Patients in the control group had higher Hb levels on admission but similar Hb levels at discharge from ICU. There was also a numerical, though not statistically significant, trend toward better preservation of the Hb at discharge in the group without transfusion (Table
2).
Patients with the blood conservation device had a significantly lower ICU and hospital mortality. While these findings must be interpreted with caution since the present study was not an RCT and mortality was not our primary or secondary end-point, they do suggest a protective effect of reduced transfusion. Indeed, blood transfusion was associated with higher mortality in both the CRIT and ABC trials [
1,
5]. Nonetheless, it should be noted that among the patients who stayed in the ICU for less than 24 hours, a larger number of patients died in the active group which may have contributed to the improved mortality in the remaining patients.
We acknowledge the limitations of our study. First, this was a before-and-after study and given the limitations of historical control study, the results of our study need to be confirmed with prospective RCT. Second, physicians and nurses were not blinded to the device. Nonetheless, we attempted to ensure equal treatment of both groups with the common restrictive transfusion strategy, which was reflected by the similar transfusion thresholds between the two periods. Third, we only included patients admitted to the medical ICU and expected to stay more than 24 hours. Although the largest volume of blood is drawn during the first 24 hours [
18], such a short study period may be insufficient to demonstrate any reduction in the PRBC transfusions. A previous study has shown that the higher mean Hb in the blood conservation group was statistically significant only after 9.5 days of ICU stay [
19]. Fourth, we excluded patients with active bleeding where transfusion practices may differ. Fifth, we used the VAMP device and it remains to be seen if our findings are applicable to other blood conservation devices.
Conclusions
Since anaemia is the main reason for transfusion in the ICU, and a blood conservation device is associated with better preservation of Hb, it is logical that use of such a device will reduce transfusion requirements. In this before-and-after study, use of a blood conservation device in the presence of a restrictive transfusion practice was indeed associated with a significant reduction in blood transfusion requirements. The significance of this finding is clear given the current worldwide shortage of PRBCs, but extends far beyond apparent cost-benefit ratio and economic savings. PRBC transfusions are associated with significant morbidity and mortality and any reduction in transfusions may eventually improve overall patient outcome. A larger prospective RCT is currently being planned.
Acknowledgements
The authors would like to thank the dedicated medical ICU nursing staff. Without their enthusiastic support this study would not be possible.
Funding for this project came from the Health Quality Improvement Fund (HQIF) from Ministry of Health (MOH), Singapore.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
AM was involved in study conception, design, securing fund, analysis and manuscript drafting. YHS was involved in institutional review board, ethics committee approval and manuscript drafting. DP was involved in data collection. CYH was involved in statistical analysis. JP and LTK were involved in manuscript drafting and PL was involved in running the project.