RBC Transfusion in Pediatric Patients Supported With Extracorporeal Membrane Oxygenation: Is There an Impact on Tissue Oxygenation?* Richard T. Fiser, MD1; Katherine Irby, MD1; Rebekah M. Ward, BS1; Xinyu Tang, PhD1; Wes McKamie, RRT, CCP2; Parthak Prodhan, MD1; Howard L. Corwin, MD3
Objective: To examine first the RBC transfusion practice in pediatric patients supported with extracorporeal membrane oxygenation and second the relationship between transfusion of RBCs and changes in mixed venous saturation (Svo2) and cerebral regional tissue oxygenation, as measured by near-infrared spectroscopy in patients supported with extracorporeal membrane oxygenation. Design: Retrospective observational study. Setting: Pediatric, cardiovascular, and neonatal ICUs of a tertiary care children’s hospital. Patients: All pediatric patients supported with extracorporeal membrane oxygenation between January 1, 2010, and December 31, 2010. Interventions: None. Measurements and Main Results: There were 45 patients supported with extracorporeal membrane oxygenation. The median (interquartile range) phlebotomy during extracorporeal membrane oxygenation was 75 mL/kg (33, 149 mL/kg). A total of 617 transfusions were administered (median, 9 per patient; range = 1–57). RBC volumes transfused during extracorporeal membrane oxygenation support were 254 mL/kg (136, 557) and 267 mL/kg (187, 393; p = 0.82) for cardiac and noncardiac patients, respectively. Subtracting the volume of RBCs used for extracorporeal membrane oxygenation circuit priming, median RBC transfusion volumes were 131 and 80 mL/kg for cardiac and noncardiac patients, respectively (p = 0.26). The cardiac surgical patients received the most RBCs (529 vs 74 mL/kg for nonsurgical cardiac patients). The median hematocrit maintained during extracorporeal membrane oxygenation support was 37%, with no difference *See also p. 895. 1 Department of Pediatrics, University of Arkansas for Medical Sciences College of Medicine, Little Rock, AR. 2 ECMO Team, Arkansas Children’s Hospital, Little Rock, AR. 3 Department of Medicine, University of Arkansas for Medical Sciences College of Medicine, Little Rock, AR. The authors have disclosed that they do not have any potential conflicts of interest. For information regarding this article, E-mail: [email protected] Copyright © 2014 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies DOI: 10.1097/PCC.0000000000000222
806
www.pccmjournal.org
between cardiac and noncardiac patients. Patients supported with extracorporeal membrane oxygenation were exposed to a median of 10.9 (range, 3–43) individual donor RBC units. Most transfusions resulted in no significant change in either Svo2 or cerebral near-infrared spectroscopy. Only 5% of transfusions administered (31/617) resulted in an increase in Svo2 of more than 5%, whereas an increase in cerebral near-infrared spectroscopy of more than 5 was only observed in 9% of transfusions (53/617). Most transfusions (73%) were administered at a time when the pretransfusion Svo2 was more than 70%. Conclusions: Patients supported with extracorporeal membrane oxygenation were exposed to large RBC transfusion volumes for treatment of mild anemia resulting from blood loss, particularly phlebotomy. In the majority of events, RBC transfusion did not significantly alter global tissue oxygenation, as assessed by changes in Svo2 and cerebral near-infrared spectroscopy. Most transfusions were administered at a time at which the patient did not appear to be oxygen delivery dependent according to global measures of tissue oxygenation. (Pediatr Crit Care Med 2014; 15:806–813) Key Words: anemia; extracorporeal life support; extracorporeal membrane oxygenation; oxygen delivery; red blood cell; transfusion
H
istorically, RBC transfusions have been viewed as a safe and effective means of treating anemia and improving oxygen delivery (Do2) to tissues. Beginning in the early 1980s, primarily driven by concerns related to the risks of transfusion-related infection, transfusion practice began to come under scrutiny. The examination of RBC transfusion risks over the last three decades has expanded to include concerns regarding age of transfused RBC, RBC storage lesions, and immunomodulation, among others, and has led to a more critical examination of RBC transfusion benefits (1). These issues are of particular importance in the critically ill. Anemia and RBC transfusion are common in critically ill adults (2, 3). It is now clear that there are little data supporting a benefit of RBC transfusion in many of the clinical situations in which RBC transfusions are routinely given (2). The few available large, November 2014 • Volume 15 • Number 9
Feature Articles
randomized clinical trials and prospective observational studies that have assessed the effectiveness of allogeneic RBC transfusion have demonstrated that a more restrictive approach to RBC transfusion results in at least equivalent patient outcomes as compared with a liberal approach and may in fact reduce morbidity and mortality rates for some critically ill patients (4). Anemia and RBC transfusion are also common in critically ill children (5). Similar to the adult critically ill, recent literature has questioned the efficacy of routine RBC transfusion in critically ill children (6–8). In a prospective, randomized, controlled trial, the Transfusion Requirements in the Pediatric Intensive Care Unit (TRIPICU) investigators demonstrated that a restrictive transfusion strategy in stable critically ill children, hemoglobin threshold of 7 g/dL, was not inferior to a liberal transfusion strategy, hemoglobin threshold 9.5 g/dL (9). The value of a higher hemoglobin concentration has similarly been questioned in the neonatal population (10). In the TRIPICU study, children randomized to the more restrictive transfusion strategy received 44% fewer RBC transfusions and demonstrated no difference in organ dysfunction or mortality with those who were transfused more liberally (9). Subgroup analyses of postsurgical (11) and postcardiac surgical (12) patients from the TRIPICU study revealed similar findings. None of these studies, however, has included pediatric patients supported with extracorporeal membrane oxygenation (ECMO). Since its early days, ECMO use has been associated with transfusion of large volumes of blood products including RBCs (13–17). Although RBC transfusions in some patients supported with ECMO are driven by bleeding complications, many RBC transfusions are ordered for the sole purpose of maintaining an arbitrary hemoglobin threshold, presumably to maintain Do2 and tissue oxygenation in these critically ill patients. Although few data exist regarding the efficacy of RBC transfusion in patients supported with ECMO, published guidelines for management of patients supported with ECMO specifically call for maintenance of a “normal” hemoglobin concentration and a hematocrit more than 40% in neonates with the aim of maximizing Do2 (18, 19). Some complications of packed RBC (PRBC) transfusion, such as transfusionrelated lung injury and hemolysis associated with the “storage lesion,” may be particularly problematic for patients supported with ECMO (20, 21). For example, patients supported with ECMO are known to be at increased risk for hemolysis, which has been associated with both greater volumes of transfused RBC and worse clinical outcomes (22, 23). Despite the rationale for RBC transfusion of maintaining adequate Do2, to date, no data exist on the impact of RBC transfusion on oxygen supply/demand balance in pediatric patients supported with ECMO. Both mixed venous saturation (Svo2) and cerebral regional tissue oxygenation, as measured by near-infrared spectroscopy (NIRS), are widely used in clinical practice as markers of global tissue Do2 (24, 25). The objective of this study is to examine first the RBC transfusion practice in a cohort of pediatric patients supported with ECMO and second the relationship between the transfusion of RBCs and changes in Svo2 and cerebral NIRS. Pediatric Critical Care Medicine
METHODS General Description This study was conducted at Arkansas Children’s Hospital, a 316-bed tertiary children’s hospital in Little Rock, AR. The study was approved by the Institutional Review Board at the University of Arkansas for Medical Sciences. Patient Population The study included all patients supported with ECMO between January 1, 2010, and December 31, 2010. Study Design The study was a retrospective review. Clinical data abstracted included demographic variables, ECMO indication and duration, hospital survival, daily hematocrit, RBC transfusion volume (including ECMO prime), and daily blood loss (phlebotomy and bleeding). Data are reported as total volume and volume/kilogram as appropriate. To evaluate the effect of RBC transfusion on tissue oxygenation, values for Svo2 and cerebral NIRS were obtained for the 2 hours before and the 2 hours following each RBC transfusion, and the mean value for each time period was analyzed and compared. A clinically significant change in Svo2 and cerebral NIRS was defined a priori as a change of more than 5%. Comparisons were made between cardiac and noncardiac patients and surgical versus nonsurgical cardiac patients. At our institution, there is no specific transfusion threshold for patients supported with ECMO. An individual attending physician determines the hematocrit threshold to be maintained by the ECMO team based on the patient’s clinical condition and the physician’s clinical judgment. Statistical Analysis Summary statistics were reported as median (25th percentile, 75th percentile) for continuous variables and percentage (frequency) for categorical variables. The distributions of continuous variables were compared between cardiac and respiratory patients using the Mann-Whitney U tests, whereas the proportions of categorical variables were compared using the chi-square test. We calculated the estimated difference in Svo2 and NIRS and their 95% CIs based on the paired pre- and posttransfusion measurements to assess whether the pre- and posttransfusion Svo2 and NIRS were equivalent. The equivalent range was predefined as from –5 to 5. p value of less than 0.05 was considered to indicate statistical significance. No adjustment was made for multiple testing. All the data were analyzed using statistical software R v3.0.1 (R Foundation for Statistical Computing, Vienna, Austria).
RESULTS Demographics Between January 1 and December 31, 2010, 45 patients were supported with ECMO. Clinical and demographic data for the study population are shown in Table 1. Twenty patients were supported primarily for cardiac indications and 25 for noncardiac indications, including respiratory failure, septic shock, www.pccmjournal.org
807
Fiser et al
Table 1.
Summary of Demographic Data Cardiac (n = 20)
n
Variable
Noncardiac (n = 25)
All Patients (n = 45)
p
2 (2, 13)
10 (2, 544)
0.003
Demographics Age (d)
45
Male
45
67 (15, 2,046) 65% (13)
56% (14)
60% (27)
0.54 0.47
Weight (kg)
45
5.5 (3.4, 36.5)
Time on extracorporeal membrane oxygenation (hr)
45
136 (95, 289)
Survived at 24 hr
45
Survived to discharge
45
3.5 (3, 3.7)
3.5 (3.3, 12.6)
0.04
164 (107, 220)
141 (107, 277)
0.63
80% (16)
92% (23)
87% (39)
0.24
60% (12)
76% (19)
69% (31)
0.25
Outcome
and extracorporeal cardiopulmonary resuscitation. Among the patients supported for a primary cardiac indication, 15 were surgical patients and five patients received ECMO support for nonsurgical cardiac indications. Median time on ECMO was 141 hours (107, 277) with no significant difference between cardiac and noncardiac patients (136 vs 164, respectively, p = 0.63) (Table 1). Survival to hospital discharge was 76% for the noncardiac patients supported with ECMO (19/25) and 60% for the cardiac group (12/20). Blood Loss and RBC Transfusion Data are shown in Table 2. A total of 617 transfusions were administered (median = 9 per patient; range = 1–57). The median (interquartile range) transfusion trigger threshold ordered by physicians was 36% (35%, 38%), with no significant difference between cardiac and noncardiac patients. Median (interquartile range) hematocrit maintained during Table 2.
ECMO support was 37% (36%, 40%) with no significant differences between cardiac and noncardiac patients. The median (interquartile range) volume of blood lost due to phlebotomy during ECMO support was 75 mL/kg (33, 149 mL/ kg),which did not differ significantly between cardiac and noncardiac patients. Cardiac patients had significantly more blood loss from chest tube and surgical site bleeding as compared with noncardiac patients (161 mL/kg [13, 336 mL/kg] vs 0 [0, 18 mL/kg]; p < 0.001). Patients received a large volume of RBCs during the course of ECMO support (254 mL/kg [136, 557] and 267 mL/kg [187, 393]; p = 0.82, for cardiac and noncardiac patients, respectively). The cardiac surgical patients received the most RBCs of any group (529 vs 74 mL/kg for nonsurgical cardiac patients). Subtracting the volume of RBCs received as part of ECMO circuit priming, median RBC transfusion volumes were 131 and 80 mL/kg for cardiac and noncardiac
Blood Loss and RBC Transfusion Volumes n
Cardiac (n = 20)
Noncardiac (n = 25)
All Patients (n = 45)
p
Average HCT transfusion threshold
45
36 (35, 38)
37 (35, 38)
36 (35, 38)
0.37
Average HCT maintained on ECMO
45
36 (34, 40)
37 (36, 41)
37 (36, 40)
0.05
Phlebotomy (mL/kg)
45
39 (20, 119)
86 (61, 149)
75 (33, 149)
0.17
Phlebotomy (mL/kg/d)
45
10.6 (2.1, 15.6)
15.3 (12.9, 16.6)
14.8 (4.8, 16.6)
0.13
Bleeding (mL/kg)
45
161 (13, 336)
0 (0, 18)
15 (0, 114)
< 0.001
Bleeding (mL/kg/d)
45
28.6 (2.3, 48.9)
0 (0, 3.7)
2.2 (0, 17.9)
< 0.001
PRBC transfusion volume without ECMO circuit primes (mL/kg)
45
131 (71, 234)
80 (51, 148)
89 (51, 212)
0.26
PRBC transfusion volume including ECMO primes (mL/kg)
45
254 (136, 557)
267 (187, 393)
267 (164, 478)
0.82
PRBC transfusion volume including ECMO circuit primes (mL/kg/d)
45
52 (29, 67)
39 (30, 52)
43 (30, 58)
0.29
Variable
ECMO = extracorporeal membrane oxygenation; PRBC = packed RBC.
808
www.pccmjournal.org
November 2014 • Volume 15 • Number 9
Feature Articles
patients, respectively (p = 0.26), with a range for all patients supported with ECMO of 51–212 mL/kg. Excluding PRBC volume for ECMO circuit primes, cardiac surgical patients received a higher PRBC transfusion volume (194 mL/kg [96, 304]) compared with nonsurgical cardiac patients (34 mL/ kg [27, 49]; p = 0.006). The RBC transfusion volume, excluding the prime volume, was greater in the noncardiac patients (80 mL/kg [51, 148]) than in the nonsurgical cardiac patients (34 mL/kg [27, 49]; p = 0.03). Including RBCs used for priming the ECMO circuit, patients supported with ECMO were exposed to a median (range) of 10.9 (3–43) individual donor RBC units during ECMO support. During the study period, the Arkansas Children’s Hospital Blood Bank released 492 individual units of PRBC for patients supported with ECMO, representing approximately 5% of all units released by the blood bank during that period. Of 617 transfusion events, data on age of transfused PRBC were available for 535 transfusions (87.7%). For these 535 transfusions, the median age of PRBC at time of transfusion was 6 days (4, 11).
Oxygen Delivery A total of 617 individual RBC transfusion events were analyzed for the association between RBC transfusion and change in Svo2 or cerebral NIRS. Most transfusions did not result in a statistically or clinically significant change in either Svo2 or cerebral NIRS (Fig. 1 and Table 3). Only 5% of the RBC transfusions administered (31/617) resulted in an increase in Svo2 of more than 5%, and an increase in cerebral NIRS of more than 5 occurred after only 9% of RBC transfusions (53/617). No relationship was found between pretransfusion hematocrit and change in either Svo2 or cerebral NIRS (Fig. 2). Most RBC transfusions (448/617, 73%) were administered when the pretransfusion Svo2 was more than 70%. When RBC transfusions were administered in situations in which the pretransfusion Svo2 was less than 70% and pretransfusion cerebral NIRS was less than 70, clinically significant changes in these variables occurred after transfusion in only a small number of instances (Table 4). No association was found between the age of transfused PRBC and changes in either Svo2 (r = 0.01; 95% CI, –0.08, 0.11; p = 0.82) or cerebral NIRS (r = –0.01; 95% CI, –0.10,
Figure 1. A, Relationship between pretransfusion and posttransfusion mixed venous oxygen saturation (Svo2). B, Relationship between pretransfusion and posttransfusion cerebral near-infrared spectroscopy (NIRS).
Table 3.
Effect of RBC Transfusion on Svo2 and Cerebral Near-Infrared Spectroscopy Change in Svo2% From Pre- to Posttransfusion
Change in Near-Infrared Spectroscopy (Absolute) From Pre- to Posttransfusion
Variable
Number of Events
Estimate (95% CI)
Estimate (95% CI)
All patients
617
0.56 (0.25, 0.88)
0.56 (0.21, 0.91)
Cardiac patients
340
0.69 (0.2, 1.17)
0.38 (–0.14, 0.9)
Surgical
314
0.71 (0.19, 1.22)
0.23 (–0.29, 0.75)
26
0.43 (–0.93, 1.79)
2.25 (–0.53, 5.03)a
Nonsurgical Respiratory patients
277
0.39 (0.04, 0.73)
0.75 (0.29, 1.21)
Neonatal
190
0.34 (–0.03, 0.70)
0.88 (0.33, 1.44)
Pediatric
87
0.57 (–0.34, 1.49)
0.47 (–0.39, 1.33)
Based on predefined equivalent range from –5 to 5, these were found to be nonequivalent.
a
Pediatric Critical Care Medicine
www.pccmjournal.org
809
Fiser et al
Figure 2. A, Relationship between pretransfusion HCT and change in transfusion mixed venous oxygen saturation (Svo2). B, Relationship between pretransfusion HCT and change in cerebral near-infrared spectroscopy (NIRS).
Table 4. Transfusion Events With Pretransfusion
Svo2 or Cerebral Near-Infrared Spectroscopy ≤ 70 and Posttransfusion Increase in Svo2 or Near-Infrared Spectroscopy ≥ 5 Variable
No
Yes
Pretransfusion Svo2 < 70 and change in Svo2 > 5
142
27
Pretransfusion NIRS < 70 and change in NIRS > 5
269
59
NIRS = near-infrared spectroscopy.
0.09; p = 0.90) (Fig. 3). For the 27 transfusion events with a pretransfusion Svo2 less than 70% and a change in Svo2 more than 5%, ΔSvo2 was not significantly different between those receiving PRBC aged less than or equal to 7 days (14/27) versus those receiving blood aged more than 7 days (8/27) (p = 0.11). Considering the 59 transfusion events with a pretransfusion NIRS less than 70 and ΔNIRS more than 5 posttransfusion, ΔNIRS was not significantly different if blood aged less than or equal to 7 days (32/59) versus more than 7 days (19/59) was received (p = 0.56).
DISCUSSION This study demonstrates a large RBC transfusion exposure in pediatric patients receiving ECMO support. Excluding the RBC transfusions for priming the ECMO circuit, the volume of these RBC transfusions was consistent with the blood loss from phlebotomy and bleeding. RBC transfusions were given to maintain the hematocrit tightly around a predefined threshold (HCT, 36%), a relatively mild degree of anemia. Furthermore, this study is the first to evaluate the relationship between RBC transfusion and global indices of tissue oxygenation (Svo2 and cerebral NIRS) in pediatric patients supported with ECMO. In the vast majority of transfusion events, RBC 810
www.pccmjournal.org
transfusion had little to no impact on these measures of global tissue oxygenation. ECMO use has been associated with transfusion of large volumes of RBCs and with exposure to a large number of donor RBC units (13–17). In a retrospective study of blood utilization in adult patients supported with ECMO, a large percentage of whom were supported for cardiogenic shock, 55% of patients received 10 or more RBC units during ECMO support, with a range of 0–42 units (26). The volume of RBCs transfused in our study, although substantial, is consistent with other studies of ECMO in pediatric patients. Smith et al (27) reported RBC transfusion volumes of 105 mL/kg/d for cardiac patients and 20 mL/kg/d for noncardiac patients; median ECMO duration was 4.6 days. Similarly, Kumar et al (28) reported RBC transfusion volume of 715 mL/kg during support with ECMO. Our findings demonstrate that in addition to the RBC volume required to prime the ECMO circuit, much of the RBC transfusion was to replace blood loss from bleeding, predominantly in cardiac patients, and phlebotomy in all patients. Bleeding, defined for this study as recorded blood loss from chest tubes, surgical/ECMO cannulation sites, or surgical drains, did occur, particularly in the cardiac surgical subset of patients and may have been the reason for some RBC transfusions. However, many of the patients supported with ECMO studied had no bleeding complications and yet received a large number of RBC transfusions. That blood loss via phlebotomy for laboratory testing contributed to the anemia experienced by these patients is consistent with a number of other studies of anemia in critically ill children and adults (5, 29, 30). However, the phlebotomy loss in our patients is particularly high compared with pediatric critically ill patients in general. The necessity, or clinical value, of the laboratory studies obtained is clearly open to question. There is no accepted RBC transfusion threshold for pediatric patients supported with ECMO. Early studies in pediatric patients supported with ECMO reported a threshold hemoglobin concentration for RBC transfusion of 15–17 g/dL November 2014 • Volume 15 • Number 9
Feature Articles
Figure 3. A, Association between age of transfused packed RBC (PRBC) and change in transfusion mixed venous oxygen saturation (Svo2). B, Association between age of transfused PRBC and change in cerebral near-infrared spectroscopy (NIRS).
(14). Guidelines published by the Extracorporeal Life Support Organization call for maintenance of a normal hematocrit and define anemia as a hematocrit less than 45% (19). Others have used HCT threshold between 35% and 45% (27, 28). Recently, Yuan et al (31), based on their experience, suggested a hematocrit threshold of 30% for transfusion. However, all these thresholds are considerably higher than both the restrictive and liberal groups in the TRIPICU trial (8). Although the restrictive threshold was equivalent to the liberal threshold in terms of outcome in the TRIPICU trial, several observational trials in pediatric patients supported with ECMO have suggested that more RBC transfusions might be associated with worse outcomes in some patients (27, 28). This raises the question of whether a more restrictive transfusion threshold might be appropriate for ECMO. In our study, most RBC transfusions were administered for treatment of anemia (maintain hemoglobin threshold of 36%) and thus presumably for the purpose of increasing global tissue Do2. However, we noted no clinically or statistically significant effect of RBC transfusion on Svo2 or cerebral NIRS in the vast majority of instances. Both Svo2 and cerebral NIRS are widely used in clinical practice as markers of global tissue Do2 (24, 25). These findings imply that most RBC transfusions were administered at a time at which the patient was not Do2 dependent, and PRBC transfusions to increase global tissue Do2 were therefore not necessary. However, this does not exclude the possibility that supply dependency could be due to tissue perfusion factors rather than inadequate blood oxygen content. Patients in this study received RBC transfusions at a median hematocrit of 36%, typically corresponding to a hemoglobin concentration of approximately 12 g/dL. At such a relatively mild degree of anemia, it is not surprising that oxygen consumption was not “supply, or delivery, dependent,” particularly in a clinical setting in which the patient’s percent oxygen saturation and cardiac output (in the case of venoarterial ECMO) were, to a large degree, controlled by the extracorporeal blood pump and membrane oxygenator. Pediatric Critical Care Medicine
Even if patients were delivery dependent, several lines of evidence question the validity of the assumption that transfusion of RBC will acutely increase Do2 at a tissue level. First, stored blood is depleted of 2,3-diphosphoglycerate (2,3-DPG), and normal levels of 2,3-DPG are not restored to transfused blood for approximately 24 hours after transfusion (32). Depletion of 2,3-DPG shifts the oxyhemoglobin dissociation curve to the left, resulting in an increased affinity of hemoglobin for oxygen at low oxygen tensions and impairing the release of oxygen to the tissues. Stored RBC also are quickly depleted of adenosine triphosphate, which results in both altered deformability and loss of integrity of the RBC membrane, which in turn can negatively impact microvascular flow and lead to early destruction of transfused RBCs (33–35). Finally, stored, transfused RBC contain small quantities of free hemoglobin that act as scavengers of endogenously produced nitric oxide, perhaps resulting in small vessel vasoconstriction and thus impairing local tissue Do2 (36, 37). Age of stored RBCs may play a role in the lack of improvement in tissue oxygenation. In both critically ill adults and children age of stored RBCs may be associated with worse tissue oxygenation and organ dysfunction (38–40). Interestingly, a small number of transfusion events in this study were associated with decreases of more than 5% in posttransfusion Svo2 and cerebral NIRS. However, no relationship was demonstrated between age of transfused blood and either Svo2 or cerebral NIRS in our study. Notably, even though most of the RBCs transfused during this study period had a relatively short storage time (median age of transfused blood = 6 d), the global markers of Do2 measured for the most part did not change after transfusion. Since most transfusions appear to have been administered at a time during which the patient was not in an Do2-dependent situation, based on global measures of Svo2 and cerebral NIRS, even transfusion of relatively fresh blood did not improve overall Do2. Our findings raise the concerning question that the majority of RBC transfusions administered to patients in this study carried no measurable benefit yet may have substantial risks. www.pccmjournal.org
811
Fiser et al
This study has several important limitations. First, it is a retrospective study of a convenience sample and thus subject to some inherent bias related to study design. More importantly, although we found no significant impact of RBC transfusion on markers of systemic Do2 in a large grouped sample, this may mask important intrapatient variability in the ability of an RBC transfusion to positively or negatively impact the oxygen supply: demand balance. At different time points during the course of ECMO support and in different pathophysiologic states, a given patient may indeed be in a “delivery-dependent” state and thus potentially benefit from transfusion. The current analysis does not fully assess the potential variability in this relationship within each patient’s ECMO course. It should also be noted that although both Svo2 measurement and monitoring of cerebral NIRS are widely used in the assessment of systemic Do2, both are subject to potential measurement error, and neither may reflect important microvascular and regional changes in Do2. Finally, we did not analyze the impact of RBC transfusion on serum lactate concentrations. Although lactate elevation may at times indicate decreased tissue Do2 and although lactate is measured commonly in patients supported with ECMO at our institution, both Svo2 and cerebral NIRS should provide earlier, more sensitive indications of global adequacy of systemic Do2 and thus seemed more appropriate targets for analysis of the impact of RBC transfusion. In conclusion, we have found that patients supported with ECMO are exposed to large RBC transfusion volumes and a large number of blood donors. In part, these transfusions are driven by blood loss, particularly phlebotomy, and are administered in response to relatively mild degrees of anemia. For most patients and in most instances, RBC transfusion to maintain the predefined HCT did not alter global measures of tissue oxygenation, as assessed by changes in Svo2 and cerebral NIRS. In such instances, RBC transfusion did not appear to have any physiologic benefit. At present, there are no published data suggesting that a higher hemoglobin is beneficial to patients supported with ECMO; however, some suggestion exists that higher RBC transfusion volumes may be associated with harm (27, 28), and there is good evidence demonstrating that a restrictive threshold is at least equivalent to a liberal threshold in critically ill pediatric patients not supported with ECMO (9). Additional studies are needed to establish the appropriate RBC transfusion threshold for pediatric patients supported with ECMO and should lend equipoise for a prospective trial of different transfusion thresholds in pediatric patients supported with ECMO. Until more data are available, a transfusion threshold no higher than the liberal threshold in TRIPICU (9.5 g/dL) should be considered.
REFERENCES
1. Hébert PC, Tinmouth A, Corwin HL: Controversies in RBC transfusion in the critically ill. Chest 2007; 131:1583–1590 2. Corwin HL, Gettinger A, Pearl RG, et al; EPO Critical Care Trials Group: Efficacy of recombinant human erythropoietin in critically ill patients: A randomized controlled trial. JAMA 2002; 288:2827–2835 3. Marik PE, Corwin HL: Efficacy of red blood cell transfusion in the critically ill: A systematic review of the literature. Crit Care Med 2008; 36:2667–2674
812
www.pccmjournal.org
4. Carson JL, Carless PA, Hebert PC: Transfusion thresholds and other strategies for guiding allogeneic red blood cell transfusion. Cochrane Database Syst Rev 2012; 4:CD002042 5. Bateman ST, Lacroix J, Boven K, et al; Pediatric Acute Lung Injury and Sepsis Investigators Network: Anemia, blood loss, and blood transfusions in North American children in the intensive care unit. Am J Respir Crit Care Med 2008; 178:26–33 6. Lacroix J, Demaret P, Tucci M: Red blood cell transfusion: Decision making in pediatric intensive care units. Semin Perinatol 2012; 36:225–231 7. Spinella PC, Dressler A, Tucci M, et al; Pediatric Acute Lung I, Sepsis Investigators Network: Survey of transfusion policies at US and Canadian children’s hospitals in 2008 and 2009. Transfusion 2010; 50:2328–2335 8. Istaphanous GK, Wheeler DS, Lisco SJ, et al: Red blood cell transfusion in critically ill children: A narrative review. Pediatr Crit Care Med 2011; 12:174–183 9. Lacroix J, Hébert PC, Hutchison JS, et al; TRIPICU Investigators; Canadian Critical Care Trials Group; Pediatric Acute Lung Injury and Sepsis Investigators Network: Transfusion strategies for patients in pediatric intensive care units. N Engl J Med 2007; 356:1609–1619 10. Whyte R, Kirpalani H: Low versus high haemoglobin concentration threshold for blood transfusion for preventing morbidity and mortality in very low birth weight infants. Cochrane Database Syst Rev 2011; 11:CD000512 11. Rouette J, Trottier H, Ducruet T, et al; Canadian Critical Care Trials Group; PALISI Network: Red blood cell transfusion threshold in postsurgical pediatric intensive care patients: A randomized clinical trial. Ann Surg 2010; 251:421–427 12. Willems A, Harrington K, Lacroix J, et al; TRIPICU investigators; Canadian Critical Care Trials Group; Pediatric Acute Lung Injury and Sepsis Investigators (PALISI) Network: Comparison of two red-cell transfusion strategies after pediatric cardiac surgery: A subgroup analysis. Crit Care Med 2010; 38:649–656 13. Butch SH, Knafl P, Oberman HA, et al: Blood utilization in adult patients undergoing extracorporeal membrane oxygenated therapy. Transfusion 1996; 36:61–63 14. Rosenberg EM, Chambers LA, Gunter JM, et al: A program to limit donor exposures to neonates undergoing extracorporeal membrane oxygenation. Pediatrics 1994; 94:341–346 15. Bjerke HS, Kelly RE Jr, Foglia RP, et al: Decreasing transfusion exposure risk during extracorporeal membrane oxygenation (ECMO). Transfus Med 1992; 2:43–49 16. Minifee PK, Daeschner CW 3rd, Griffin MP, et al: Decreasing blood donor exposure in neonates on extracorporeal membrane oxygenation. J Pediatr Surg 1990; 25:38–42 17. Green TP, Payne NR, Steinhorn RH: Determinants of blood product use during extracorporeal membrane oxygenation. Transfusion 1990; 30:289–290 18. Van Meurs KPHS, Sheehan AM: ECMO for neonatal respiratory failure. In: ECMO: Extracorporeal Cardiopulmonary Support in Critical Care. Van Meurs KP, Lally KP, Peek G, et al (Eds). Third Edition. Ann Arbor, MI, Extracorporeal Life Support Organization, 2005, pp 273–296 19. Extracorporeal Life Support Organization: General Guidelines for All ECLS Cases. Available at: http://www.elsonet.org/index. php?option=com_phocadownload&view=category&id=4&Ite mid=627. Accessed September 28, 2013 20. Weber LL, Roberts LD, Sweeney JD: Residual plasma in red blood cells and transfusion-related acute lung injury. Transfusion 2014 Apr 25 [Epub ahead of print] 21. Lieberman L, Petraszko T, Yi QL, et al: Transfusion-related lung injury in children: A case series and review of the literature. Transfusion 2014; 54:57–64 22. Lou S, MacLaren G, Best D, et al: Hemolysis in pediatric patients receiving centrifugal-pump extracorporeal membrane oxygenation: Prevalence, risk factors, and outcomes. Crit Care Med 2014; 42:1213–1220 23. Byrnes J, McKamie W, Swearingen C, et al: Hemolysis during cardiac extracorporeal membrane oxygenation: A case-control comparison of roller pumps and centrifugal pumps in a pediatric population. ASAIO J 2011; 57:456–461 November 2014 • Volume 15 • Number 9
Feature Articles 24. Dellinger RP, Levy MM, Rhodes A, et al: Surviving sepsis campaign: International guidelines for management of severe sepsis and septic shock, 2012. Intensive Care Med 2013; 39:165–228 25. Tyree K, Tyree M, DiGeronimo R: Correlation of brain tissue oxygen tension with cerebral near-infrared spectroscopy and mixed venous oxygen saturation during extracorporeal membrane oxygenation. Perfusion 2009; 24:325–331 26. Ang AL, Teo D, Lim CH, et al: Blood transfusion requirements and independent predictors of increased transfusion requirements among adult patients on extracorporeal membrane oxygenation—a single centre experience. Vox Sang 2009; 96:34–43 27. Smith A, Hardison D, Bridges B, et al: Red blood cell transfusion volume and mortality among patients receiving extracorporeal membrane oxygenation. Perfusion 2013; 28:54–60 28. Kumar TK, Zurakowski D, Dalton H, et al: Extracorporeal membrane oxygenation in postcardiotomy patients: Factors influencing outcome. J Thorac Cardiovasc Surg 2010; 140:330–336.e332 29. Burnum JF: Medical vampires. N Engl J Med 1986; 314:1250–1251 30. Corwin HL, Gettinger A, Pearl RG, et al: The CRIT Study: Anemia and blood transfusion in the critically ill–current clinical practice in the United States. Crit Care Med 2004; 32:39–52 31. Yuan S, Tsukahara E, De La Cruz K, et al: How we provide transfusion support for neonatal and pediatric patients on extracorporeal membrane oxygenation. Transfusion 2013; 53:1157–1165 32. Klein HG, Spahn DR, Carson JL: Red blood cell transfusion in clinical practice. Lancet 2007; 370:415–426
Pediatric Critical Care Medicine
33. Luten M, Roerdinkholder-Stoelwinder B, Schaap NP, et al: Survival of red blood cells after transfusion: A comparison between red cells concentrates of different storage periods. Transfusion 2008; 48:1478–1485 34. Bosman GJ, Werre JM, Willekens FL, et al: Erythrocyte ageing in vivo and in vitro: Structural aspects and implications for transfusion. Transfus Med 2008; 18:335–347 35. Tinmouth A, Fergusson D, Yee IC, et al; ABLE Investigators; Canadian Critical Care Trials Group: Clinical consequences of red cell storage in the critically ill. Transfusion 2006; 46:2014–2027 36. Doctor A, Platt R, Sheram ML, et al: Hemoglobin conformation couples erythrocyte S-nitrosothiol content to O2 gradients. Proc Natl Acad Sci U S A 2005; 102:5709–5714 37. Reiter CD, Wang X, Tanus-Santos JE, et al: Cell-free hemoglobin limits nitric oxide bioavailability in sickle-cell disease. Nat Med 2002; 8:1383–1389 38. Marik PE, Sibbald WJ: Effect of stored-blood transfusion on oxygen delivery in patients with sepsis. JAMA 1993; 269:3024–3029 39. Karam O, Tucci M, Bateman ST, et al: Association between length of storage of red blood cell units and outcome of critically ill children: A prospective observational study. Crit Care 2010; 14:R57 40. Gauvin F, Spinella PC, Lacroix J, et al; Canadian Critical Care Trials Group and the Pediatric Acute Lung Injury and Sepsis Investigators (PALISI) Network: Association between length of storage of transfused red blood cells and multiple organ dysfunction syndrome in pediatric intensive care patients. Transfusion 2010; 50:1902–1913
www.pccmjournal.org
813
Objective: To examine first the RBC transfusion practice in pediatric patients supported with extracorporeal membrane oxygenation and second the relationship between transfusion of RBCs and changes in mixed venous saturation (Svo2) and cerebral regional tissue oxygenation, as measured by near-infrared spectroscopy in patients supported with extracorporeal membrane oxygenation. Design: Retrospective observational study. Setting: Pediatric, cardiovascular, and neonatal ICUs of a tertiary care children’s hospital. Patients: All pediatric patients supported with extracorporeal membrane oxygenation between January 1, 2010, and December 31, 2010. Interventions: None. Measurements and Main Results: There were 45 patients supported with extracorporeal membrane oxygenation. The median (interquartile range) phlebotomy during extracorporeal membrane oxygenation was 75 mL/kg (33, 149 mL/kg). A total of 617 transfusions were administered (median, 9 per patient; range = 1–57). RBC volumes transfused during extracorporeal membrane oxygenation support were 254 mL/kg (136, 557) and 267 mL/kg (187, 393; p = 0.82) for cardiac and noncardiac patients, respectively. Subtracting the volume of RBCs used for extracorporeal membrane oxygenation circuit priming, median RBC transfusion volumes were 131 and 80 mL/kg for cardiac and noncardiac patients, respectively (p = 0.26). The cardiac surgical patients received the most RBCs (529 vs 74 mL/kg for nonsurgical cardiac patients). The median hematocrit maintained during extracorporeal membrane oxygenation support was 37%, with no difference *See also p. 895. 1 Department of Pediatrics, University of Arkansas for Medical Sciences College of Medicine, Little Rock, AR. 2 ECMO Team, Arkansas Children’s Hospital, Little Rock, AR. 3 Department of Medicine, University of Arkansas for Medical Sciences College of Medicine, Little Rock, AR. The authors have disclosed that they do not have any potential conflicts of interest. For information regarding this article, E-mail: [email protected] Copyright © 2014 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies DOI: 10.1097/PCC.0000000000000222
806
www.pccmjournal.org
between cardiac and noncardiac patients. Patients supported with extracorporeal membrane oxygenation were exposed to a median of 10.9 (range, 3–43) individual donor RBC units. Most transfusions resulted in no significant change in either Svo2 or cerebral near-infrared spectroscopy. Only 5% of transfusions administered (31/617) resulted in an increase in Svo2 of more than 5%, whereas an increase in cerebral near-infrared spectroscopy of more than 5 was only observed in 9% of transfusions (53/617). Most transfusions (73%) were administered at a time when the pretransfusion Svo2 was more than 70%. Conclusions: Patients supported with extracorporeal membrane oxygenation were exposed to large RBC transfusion volumes for treatment of mild anemia resulting from blood loss, particularly phlebotomy. In the majority of events, RBC transfusion did not significantly alter global tissue oxygenation, as assessed by changes in Svo2 and cerebral near-infrared spectroscopy. Most transfusions were administered at a time at which the patient did not appear to be oxygen delivery dependent according to global measures of tissue oxygenation. (Pediatr Crit Care Med 2014; 15:806–813) Key Words: anemia; extracorporeal life support; extracorporeal membrane oxygenation; oxygen delivery; red blood cell; transfusion
H
istorically, RBC transfusions have been viewed as a safe and effective means of treating anemia and improving oxygen delivery (Do2) to tissues. Beginning in the early 1980s, primarily driven by concerns related to the risks of transfusion-related infection, transfusion practice began to come under scrutiny. The examination of RBC transfusion risks over the last three decades has expanded to include concerns regarding age of transfused RBC, RBC storage lesions, and immunomodulation, among others, and has led to a more critical examination of RBC transfusion benefits (1). These issues are of particular importance in the critically ill. Anemia and RBC transfusion are common in critically ill adults (2, 3). It is now clear that there are little data supporting a benefit of RBC transfusion in many of the clinical situations in which RBC transfusions are routinely given (2). The few available large, November 2014 • Volume 15 • Number 9
Feature Articles
randomized clinical trials and prospective observational studies that have assessed the effectiveness of allogeneic RBC transfusion have demonstrated that a more restrictive approach to RBC transfusion results in at least equivalent patient outcomes as compared with a liberal approach and may in fact reduce morbidity and mortality rates for some critically ill patients (4). Anemia and RBC transfusion are also common in critically ill children (5). Similar to the adult critically ill, recent literature has questioned the efficacy of routine RBC transfusion in critically ill children (6–8). In a prospective, randomized, controlled trial, the Transfusion Requirements in the Pediatric Intensive Care Unit (TRIPICU) investigators demonstrated that a restrictive transfusion strategy in stable critically ill children, hemoglobin threshold of 7 g/dL, was not inferior to a liberal transfusion strategy, hemoglobin threshold 9.5 g/dL (9). The value of a higher hemoglobin concentration has similarly been questioned in the neonatal population (10). In the TRIPICU study, children randomized to the more restrictive transfusion strategy received 44% fewer RBC transfusions and demonstrated no difference in organ dysfunction or mortality with those who were transfused more liberally (9). Subgroup analyses of postsurgical (11) and postcardiac surgical (12) patients from the TRIPICU study revealed similar findings. None of these studies, however, has included pediatric patients supported with extracorporeal membrane oxygenation (ECMO). Since its early days, ECMO use has been associated with transfusion of large volumes of blood products including RBCs (13–17). Although RBC transfusions in some patients supported with ECMO are driven by bleeding complications, many RBC transfusions are ordered for the sole purpose of maintaining an arbitrary hemoglobin threshold, presumably to maintain Do2 and tissue oxygenation in these critically ill patients. Although few data exist regarding the efficacy of RBC transfusion in patients supported with ECMO, published guidelines for management of patients supported with ECMO specifically call for maintenance of a “normal” hemoglobin concentration and a hematocrit more than 40% in neonates with the aim of maximizing Do2 (18, 19). Some complications of packed RBC (PRBC) transfusion, such as transfusionrelated lung injury and hemolysis associated with the “storage lesion,” may be particularly problematic for patients supported with ECMO (20, 21). For example, patients supported with ECMO are known to be at increased risk for hemolysis, which has been associated with both greater volumes of transfused RBC and worse clinical outcomes (22, 23). Despite the rationale for RBC transfusion of maintaining adequate Do2, to date, no data exist on the impact of RBC transfusion on oxygen supply/demand balance in pediatric patients supported with ECMO. Both mixed venous saturation (Svo2) and cerebral regional tissue oxygenation, as measured by near-infrared spectroscopy (NIRS), are widely used in clinical practice as markers of global tissue Do2 (24, 25). The objective of this study is to examine first the RBC transfusion practice in a cohort of pediatric patients supported with ECMO and second the relationship between the transfusion of RBCs and changes in Svo2 and cerebral NIRS. Pediatric Critical Care Medicine
METHODS General Description This study was conducted at Arkansas Children’s Hospital, a 316-bed tertiary children’s hospital in Little Rock, AR. The study was approved by the Institutional Review Board at the University of Arkansas for Medical Sciences. Patient Population The study included all patients supported with ECMO between January 1, 2010, and December 31, 2010. Study Design The study was a retrospective review. Clinical data abstracted included demographic variables, ECMO indication and duration, hospital survival, daily hematocrit, RBC transfusion volume (including ECMO prime), and daily blood loss (phlebotomy and bleeding). Data are reported as total volume and volume/kilogram as appropriate. To evaluate the effect of RBC transfusion on tissue oxygenation, values for Svo2 and cerebral NIRS were obtained for the 2 hours before and the 2 hours following each RBC transfusion, and the mean value for each time period was analyzed and compared. A clinically significant change in Svo2 and cerebral NIRS was defined a priori as a change of more than 5%. Comparisons were made between cardiac and noncardiac patients and surgical versus nonsurgical cardiac patients. At our institution, there is no specific transfusion threshold for patients supported with ECMO. An individual attending physician determines the hematocrit threshold to be maintained by the ECMO team based on the patient’s clinical condition and the physician’s clinical judgment. Statistical Analysis Summary statistics were reported as median (25th percentile, 75th percentile) for continuous variables and percentage (frequency) for categorical variables. The distributions of continuous variables were compared between cardiac and respiratory patients using the Mann-Whitney U tests, whereas the proportions of categorical variables were compared using the chi-square test. We calculated the estimated difference in Svo2 and NIRS and their 95% CIs based on the paired pre- and posttransfusion measurements to assess whether the pre- and posttransfusion Svo2 and NIRS were equivalent. The equivalent range was predefined as from –5 to 5. p value of less than 0.05 was considered to indicate statistical significance. No adjustment was made for multiple testing. All the data were analyzed using statistical software R v3.0.1 (R Foundation for Statistical Computing, Vienna, Austria).
RESULTS Demographics Between January 1 and December 31, 2010, 45 patients were supported with ECMO. Clinical and demographic data for the study population are shown in Table 1. Twenty patients were supported primarily for cardiac indications and 25 for noncardiac indications, including respiratory failure, septic shock, www.pccmjournal.org
807
Fiser et al
Table 1.
Summary of Demographic Data Cardiac (n = 20)
n
Variable
Noncardiac (n = 25)
All Patients (n = 45)
p
2 (2, 13)
10 (2, 544)
0.003
Demographics Age (d)
45
Male
45
67 (15, 2,046) 65% (13)
56% (14)
60% (27)
0.54 0.47
Weight (kg)
45
5.5 (3.4, 36.5)
Time on extracorporeal membrane oxygenation (hr)
45
136 (95, 289)
Survived at 24 hr
45
Survived to discharge
45
3.5 (3, 3.7)
3.5 (3.3, 12.6)
0.04
164 (107, 220)
141 (107, 277)
0.63
80% (16)
92% (23)
87% (39)
0.24
60% (12)
76% (19)
69% (31)
0.25
Outcome
and extracorporeal cardiopulmonary resuscitation. Among the patients supported for a primary cardiac indication, 15 were surgical patients and five patients received ECMO support for nonsurgical cardiac indications. Median time on ECMO was 141 hours (107, 277) with no significant difference between cardiac and noncardiac patients (136 vs 164, respectively, p = 0.63) (Table 1). Survival to hospital discharge was 76% for the noncardiac patients supported with ECMO (19/25) and 60% for the cardiac group (12/20). Blood Loss and RBC Transfusion Data are shown in Table 2. A total of 617 transfusions were administered (median = 9 per patient; range = 1–57). The median (interquartile range) transfusion trigger threshold ordered by physicians was 36% (35%, 38%), with no significant difference between cardiac and noncardiac patients. Median (interquartile range) hematocrit maintained during Table 2.
ECMO support was 37% (36%, 40%) with no significant differences between cardiac and noncardiac patients. The median (interquartile range) volume of blood lost due to phlebotomy during ECMO support was 75 mL/kg (33, 149 mL/ kg),which did not differ significantly between cardiac and noncardiac patients. Cardiac patients had significantly more blood loss from chest tube and surgical site bleeding as compared with noncardiac patients (161 mL/kg [13, 336 mL/kg] vs 0 [0, 18 mL/kg]; p < 0.001). Patients received a large volume of RBCs during the course of ECMO support (254 mL/kg [136, 557] and 267 mL/kg [187, 393]; p = 0.82, for cardiac and noncardiac patients, respectively). The cardiac surgical patients received the most RBCs of any group (529 vs 74 mL/kg for nonsurgical cardiac patients). Subtracting the volume of RBCs received as part of ECMO circuit priming, median RBC transfusion volumes were 131 and 80 mL/kg for cardiac and noncardiac
Blood Loss and RBC Transfusion Volumes n
Cardiac (n = 20)
Noncardiac (n = 25)
All Patients (n = 45)
p
Average HCT transfusion threshold
45
36 (35, 38)
37 (35, 38)
36 (35, 38)
0.37
Average HCT maintained on ECMO
45
36 (34, 40)
37 (36, 41)
37 (36, 40)
0.05
Phlebotomy (mL/kg)
45
39 (20, 119)
86 (61, 149)
75 (33, 149)
0.17
Phlebotomy (mL/kg/d)
45
10.6 (2.1, 15.6)
15.3 (12.9, 16.6)
14.8 (4.8, 16.6)
0.13
Bleeding (mL/kg)
45
161 (13, 336)
0 (0, 18)
15 (0, 114)
< 0.001
Bleeding (mL/kg/d)
45
28.6 (2.3, 48.9)
0 (0, 3.7)
2.2 (0, 17.9)
< 0.001
PRBC transfusion volume without ECMO circuit primes (mL/kg)
45
131 (71, 234)
80 (51, 148)
89 (51, 212)
0.26
PRBC transfusion volume including ECMO primes (mL/kg)
45
254 (136, 557)
267 (187, 393)
267 (164, 478)
0.82
PRBC transfusion volume including ECMO circuit primes (mL/kg/d)
45
52 (29, 67)
39 (30, 52)
43 (30, 58)
0.29
Variable
ECMO = extracorporeal membrane oxygenation; PRBC = packed RBC.
808
www.pccmjournal.org
November 2014 • Volume 15 • Number 9
Feature Articles
patients, respectively (p = 0.26), with a range for all patients supported with ECMO of 51–212 mL/kg. Excluding PRBC volume for ECMO circuit primes, cardiac surgical patients received a higher PRBC transfusion volume (194 mL/kg [96, 304]) compared with nonsurgical cardiac patients (34 mL/ kg [27, 49]; p = 0.006). The RBC transfusion volume, excluding the prime volume, was greater in the noncardiac patients (80 mL/kg [51, 148]) than in the nonsurgical cardiac patients (34 mL/kg [27, 49]; p = 0.03). Including RBCs used for priming the ECMO circuit, patients supported with ECMO were exposed to a median (range) of 10.9 (3–43) individual donor RBC units during ECMO support. During the study period, the Arkansas Children’s Hospital Blood Bank released 492 individual units of PRBC for patients supported with ECMO, representing approximately 5% of all units released by the blood bank during that period. Of 617 transfusion events, data on age of transfused PRBC were available for 535 transfusions (87.7%). For these 535 transfusions, the median age of PRBC at time of transfusion was 6 days (4, 11).
Oxygen Delivery A total of 617 individual RBC transfusion events were analyzed for the association between RBC transfusion and change in Svo2 or cerebral NIRS. Most transfusions did not result in a statistically or clinically significant change in either Svo2 or cerebral NIRS (Fig. 1 and Table 3). Only 5% of the RBC transfusions administered (31/617) resulted in an increase in Svo2 of more than 5%, and an increase in cerebral NIRS of more than 5 occurred after only 9% of RBC transfusions (53/617). No relationship was found between pretransfusion hematocrit and change in either Svo2 or cerebral NIRS (Fig. 2). Most RBC transfusions (448/617, 73%) were administered when the pretransfusion Svo2 was more than 70%. When RBC transfusions were administered in situations in which the pretransfusion Svo2 was less than 70% and pretransfusion cerebral NIRS was less than 70, clinically significant changes in these variables occurred after transfusion in only a small number of instances (Table 4). No association was found between the age of transfused PRBC and changes in either Svo2 (r = 0.01; 95% CI, –0.08, 0.11; p = 0.82) or cerebral NIRS (r = –0.01; 95% CI, –0.10,
Figure 1. A, Relationship between pretransfusion and posttransfusion mixed venous oxygen saturation (Svo2). B, Relationship between pretransfusion and posttransfusion cerebral near-infrared spectroscopy (NIRS).
Table 3.
Effect of RBC Transfusion on Svo2 and Cerebral Near-Infrared Spectroscopy Change in Svo2% From Pre- to Posttransfusion
Change in Near-Infrared Spectroscopy (Absolute) From Pre- to Posttransfusion
Variable
Number of Events
Estimate (95% CI)
Estimate (95% CI)
All patients
617
0.56 (0.25, 0.88)
0.56 (0.21, 0.91)
Cardiac patients
340
0.69 (0.2, 1.17)
0.38 (–0.14, 0.9)
Surgical
314
0.71 (0.19, 1.22)
0.23 (–0.29, 0.75)
26
0.43 (–0.93, 1.79)
2.25 (–0.53, 5.03)a
Nonsurgical Respiratory patients
277
0.39 (0.04, 0.73)
0.75 (0.29, 1.21)
Neonatal
190
0.34 (–0.03, 0.70)
0.88 (0.33, 1.44)
Pediatric
87
0.57 (–0.34, 1.49)
0.47 (–0.39, 1.33)
Based on predefined equivalent range from –5 to 5, these were found to be nonequivalent.
a
Pediatric Critical Care Medicine
www.pccmjournal.org
809
Fiser et al
Figure 2. A, Relationship between pretransfusion HCT and change in transfusion mixed venous oxygen saturation (Svo2). B, Relationship between pretransfusion HCT and change in cerebral near-infrared spectroscopy (NIRS).
Table 4. Transfusion Events With Pretransfusion
Svo2 or Cerebral Near-Infrared Spectroscopy ≤ 70 and Posttransfusion Increase in Svo2 or Near-Infrared Spectroscopy ≥ 5 Variable
No
Yes
Pretransfusion Svo2 < 70 and change in Svo2 > 5
142
27
Pretransfusion NIRS < 70 and change in NIRS > 5
269
59
NIRS = near-infrared spectroscopy.
0.09; p = 0.90) (Fig. 3). For the 27 transfusion events with a pretransfusion Svo2 less than 70% and a change in Svo2 more than 5%, ΔSvo2 was not significantly different between those receiving PRBC aged less than or equal to 7 days (14/27) versus those receiving blood aged more than 7 days (8/27) (p = 0.11). Considering the 59 transfusion events with a pretransfusion NIRS less than 70 and ΔNIRS more than 5 posttransfusion, ΔNIRS was not significantly different if blood aged less than or equal to 7 days (32/59) versus more than 7 days (19/59) was received (p = 0.56).
DISCUSSION This study demonstrates a large RBC transfusion exposure in pediatric patients receiving ECMO support. Excluding the RBC transfusions for priming the ECMO circuit, the volume of these RBC transfusions was consistent with the blood loss from phlebotomy and bleeding. RBC transfusions were given to maintain the hematocrit tightly around a predefined threshold (HCT, 36%), a relatively mild degree of anemia. Furthermore, this study is the first to evaluate the relationship between RBC transfusion and global indices of tissue oxygenation (Svo2 and cerebral NIRS) in pediatric patients supported with ECMO. In the vast majority of transfusion events, RBC 810
www.pccmjournal.org
transfusion had little to no impact on these measures of global tissue oxygenation. ECMO use has been associated with transfusion of large volumes of RBCs and with exposure to a large number of donor RBC units (13–17). In a retrospective study of blood utilization in adult patients supported with ECMO, a large percentage of whom were supported for cardiogenic shock, 55% of patients received 10 or more RBC units during ECMO support, with a range of 0–42 units (26). The volume of RBCs transfused in our study, although substantial, is consistent with other studies of ECMO in pediatric patients. Smith et al (27) reported RBC transfusion volumes of 105 mL/kg/d for cardiac patients and 20 mL/kg/d for noncardiac patients; median ECMO duration was 4.6 days. Similarly, Kumar et al (28) reported RBC transfusion volume of 715 mL/kg during support with ECMO. Our findings demonstrate that in addition to the RBC volume required to prime the ECMO circuit, much of the RBC transfusion was to replace blood loss from bleeding, predominantly in cardiac patients, and phlebotomy in all patients. Bleeding, defined for this study as recorded blood loss from chest tubes, surgical/ECMO cannulation sites, or surgical drains, did occur, particularly in the cardiac surgical subset of patients and may have been the reason for some RBC transfusions. However, many of the patients supported with ECMO studied had no bleeding complications and yet received a large number of RBC transfusions. That blood loss via phlebotomy for laboratory testing contributed to the anemia experienced by these patients is consistent with a number of other studies of anemia in critically ill children and adults (5, 29, 30). However, the phlebotomy loss in our patients is particularly high compared with pediatric critically ill patients in general. The necessity, or clinical value, of the laboratory studies obtained is clearly open to question. There is no accepted RBC transfusion threshold for pediatric patients supported with ECMO. Early studies in pediatric patients supported with ECMO reported a threshold hemoglobin concentration for RBC transfusion of 15–17 g/dL November 2014 • Volume 15 • Number 9
Feature Articles
Figure 3. A, Association between age of transfused packed RBC (PRBC) and change in transfusion mixed venous oxygen saturation (Svo2). B, Association between age of transfused PRBC and change in cerebral near-infrared spectroscopy (NIRS).
(14). Guidelines published by the Extracorporeal Life Support Organization call for maintenance of a normal hematocrit and define anemia as a hematocrit less than 45% (19). Others have used HCT threshold between 35% and 45% (27, 28). Recently, Yuan et al (31), based on their experience, suggested a hematocrit threshold of 30% for transfusion. However, all these thresholds are considerably higher than both the restrictive and liberal groups in the TRIPICU trial (8). Although the restrictive threshold was equivalent to the liberal threshold in terms of outcome in the TRIPICU trial, several observational trials in pediatric patients supported with ECMO have suggested that more RBC transfusions might be associated with worse outcomes in some patients (27, 28). This raises the question of whether a more restrictive transfusion threshold might be appropriate for ECMO. In our study, most RBC transfusions were administered for treatment of anemia (maintain hemoglobin threshold of 36%) and thus presumably for the purpose of increasing global tissue Do2. However, we noted no clinically or statistically significant effect of RBC transfusion on Svo2 or cerebral NIRS in the vast majority of instances. Both Svo2 and cerebral NIRS are widely used in clinical practice as markers of global tissue Do2 (24, 25). These findings imply that most RBC transfusions were administered at a time at which the patient was not Do2 dependent, and PRBC transfusions to increase global tissue Do2 were therefore not necessary. However, this does not exclude the possibility that supply dependency could be due to tissue perfusion factors rather than inadequate blood oxygen content. Patients in this study received RBC transfusions at a median hematocrit of 36%, typically corresponding to a hemoglobin concentration of approximately 12 g/dL. At such a relatively mild degree of anemia, it is not surprising that oxygen consumption was not “supply, or delivery, dependent,” particularly in a clinical setting in which the patient’s percent oxygen saturation and cardiac output (in the case of venoarterial ECMO) were, to a large degree, controlled by the extracorporeal blood pump and membrane oxygenator. Pediatric Critical Care Medicine
Even if patients were delivery dependent, several lines of evidence question the validity of the assumption that transfusion of RBC will acutely increase Do2 at a tissue level. First, stored blood is depleted of 2,3-diphosphoglycerate (2,3-DPG), and normal levels of 2,3-DPG are not restored to transfused blood for approximately 24 hours after transfusion (32). Depletion of 2,3-DPG shifts the oxyhemoglobin dissociation curve to the left, resulting in an increased affinity of hemoglobin for oxygen at low oxygen tensions and impairing the release of oxygen to the tissues. Stored RBC also are quickly depleted of adenosine triphosphate, which results in both altered deformability and loss of integrity of the RBC membrane, which in turn can negatively impact microvascular flow and lead to early destruction of transfused RBCs (33–35). Finally, stored, transfused RBC contain small quantities of free hemoglobin that act as scavengers of endogenously produced nitric oxide, perhaps resulting in small vessel vasoconstriction and thus impairing local tissue Do2 (36, 37). Age of stored RBCs may play a role in the lack of improvement in tissue oxygenation. In both critically ill adults and children age of stored RBCs may be associated with worse tissue oxygenation and organ dysfunction (38–40). Interestingly, a small number of transfusion events in this study were associated with decreases of more than 5% in posttransfusion Svo2 and cerebral NIRS. However, no relationship was demonstrated between age of transfused blood and either Svo2 or cerebral NIRS in our study. Notably, even though most of the RBCs transfused during this study period had a relatively short storage time (median age of transfused blood = 6 d), the global markers of Do2 measured for the most part did not change after transfusion. Since most transfusions appear to have been administered at a time during which the patient was not in an Do2-dependent situation, based on global measures of Svo2 and cerebral NIRS, even transfusion of relatively fresh blood did not improve overall Do2. Our findings raise the concerning question that the majority of RBC transfusions administered to patients in this study carried no measurable benefit yet may have substantial risks. www.pccmjournal.org
811
Fiser et al
This study has several important limitations. First, it is a retrospective study of a convenience sample and thus subject to some inherent bias related to study design. More importantly, although we found no significant impact of RBC transfusion on markers of systemic Do2 in a large grouped sample, this may mask important intrapatient variability in the ability of an RBC transfusion to positively or negatively impact the oxygen supply: demand balance. At different time points during the course of ECMO support and in different pathophysiologic states, a given patient may indeed be in a “delivery-dependent” state and thus potentially benefit from transfusion. The current analysis does not fully assess the potential variability in this relationship within each patient’s ECMO course. It should also be noted that although both Svo2 measurement and monitoring of cerebral NIRS are widely used in the assessment of systemic Do2, both are subject to potential measurement error, and neither may reflect important microvascular and regional changes in Do2. Finally, we did not analyze the impact of RBC transfusion on serum lactate concentrations. Although lactate elevation may at times indicate decreased tissue Do2 and although lactate is measured commonly in patients supported with ECMO at our institution, both Svo2 and cerebral NIRS should provide earlier, more sensitive indications of global adequacy of systemic Do2 and thus seemed more appropriate targets for analysis of the impact of RBC transfusion. In conclusion, we have found that patients supported with ECMO are exposed to large RBC transfusion volumes and a large number of blood donors. In part, these transfusions are driven by blood loss, particularly phlebotomy, and are administered in response to relatively mild degrees of anemia. For most patients and in most instances, RBC transfusion to maintain the predefined HCT did not alter global measures of tissue oxygenation, as assessed by changes in Svo2 and cerebral NIRS. In such instances, RBC transfusion did not appear to have any physiologic benefit. At present, there are no published data suggesting that a higher hemoglobin is beneficial to patients supported with ECMO; however, some suggestion exists that higher RBC transfusion volumes may be associated with harm (27, 28), and there is good evidence demonstrating that a restrictive threshold is at least equivalent to a liberal threshold in critically ill pediatric patients not supported with ECMO (9). Additional studies are needed to establish the appropriate RBC transfusion threshold for pediatric patients supported with ECMO and should lend equipoise for a prospective trial of different transfusion thresholds in pediatric patients supported with ECMO. Until more data are available, a transfusion threshold no higher than the liberal threshold in TRIPICU (9.5 g/dL) should be considered.
REFERENCES
1. Hébert PC, Tinmouth A, Corwin HL: Controversies in RBC transfusion in the critically ill. Chest 2007; 131:1583–1590 2. Corwin HL, Gettinger A, Pearl RG, et al; EPO Critical Care Trials Group: Efficacy of recombinant human erythropoietin in critically ill patients: A randomized controlled trial. JAMA 2002; 288:2827–2835 3. Marik PE, Corwin HL: Efficacy of red blood cell transfusion in the critically ill: A systematic review of the literature. Crit Care Med 2008; 36:2667–2674
812
www.pccmjournal.org
4. Carson JL, Carless PA, Hebert PC: Transfusion thresholds and other strategies for guiding allogeneic red blood cell transfusion. Cochrane Database Syst Rev 2012; 4:CD002042 5. Bateman ST, Lacroix J, Boven K, et al; Pediatric Acute Lung Injury and Sepsis Investigators Network: Anemia, blood loss, and blood transfusions in North American children in the intensive care unit. Am J Respir Crit Care Med 2008; 178:26–33 6. Lacroix J, Demaret P, Tucci M: Red blood cell transfusion: Decision making in pediatric intensive care units. Semin Perinatol 2012; 36:225–231 7. Spinella PC, Dressler A, Tucci M, et al; Pediatric Acute Lung I, Sepsis Investigators Network: Survey of transfusion policies at US and Canadian children’s hospitals in 2008 and 2009. Transfusion 2010; 50:2328–2335 8. Istaphanous GK, Wheeler DS, Lisco SJ, et al: Red blood cell transfusion in critically ill children: A narrative review. Pediatr Crit Care Med 2011; 12:174–183 9. Lacroix J, Hébert PC, Hutchison JS, et al; TRIPICU Investigators; Canadian Critical Care Trials Group; Pediatric Acute Lung Injury and Sepsis Investigators Network: Transfusion strategies for patients in pediatric intensive care units. N Engl J Med 2007; 356:1609–1619 10. Whyte R, Kirpalani H: Low versus high haemoglobin concentration threshold for blood transfusion for preventing morbidity and mortality in very low birth weight infants. Cochrane Database Syst Rev 2011; 11:CD000512 11. Rouette J, Trottier H, Ducruet T, et al; Canadian Critical Care Trials Group; PALISI Network: Red blood cell transfusion threshold in postsurgical pediatric intensive care patients: A randomized clinical trial. Ann Surg 2010; 251:421–427 12. Willems A, Harrington K, Lacroix J, et al; TRIPICU investigators; Canadian Critical Care Trials Group; Pediatric Acute Lung Injury and Sepsis Investigators (PALISI) Network: Comparison of two red-cell transfusion strategies after pediatric cardiac surgery: A subgroup analysis. Crit Care Med 2010; 38:649–656 13. Butch SH, Knafl P, Oberman HA, et al: Blood utilization in adult patients undergoing extracorporeal membrane oxygenated therapy. Transfusion 1996; 36:61–63 14. Rosenberg EM, Chambers LA, Gunter JM, et al: A program to limit donor exposures to neonates undergoing extracorporeal membrane oxygenation. Pediatrics 1994; 94:341–346 15. Bjerke HS, Kelly RE Jr, Foglia RP, et al: Decreasing transfusion exposure risk during extracorporeal membrane oxygenation (ECMO). Transfus Med 1992; 2:43–49 16. Minifee PK, Daeschner CW 3rd, Griffin MP, et al: Decreasing blood donor exposure in neonates on extracorporeal membrane oxygenation. J Pediatr Surg 1990; 25:38–42 17. Green TP, Payne NR, Steinhorn RH: Determinants of blood product use during extracorporeal membrane oxygenation. Transfusion 1990; 30:289–290 18. Van Meurs KPHS, Sheehan AM: ECMO for neonatal respiratory failure. In: ECMO: Extracorporeal Cardiopulmonary Support in Critical Care. Van Meurs KP, Lally KP, Peek G, et al (Eds). Third Edition. Ann Arbor, MI, Extracorporeal Life Support Organization, 2005, pp 273–296 19. Extracorporeal Life Support Organization: General Guidelines for All ECLS Cases. Available at: http://www.elsonet.org/index. php?option=com_phocadownload&view=category&id=4&Ite mid=627. Accessed September 28, 2013 20. Weber LL, Roberts LD, Sweeney JD: Residual plasma in red blood cells and transfusion-related acute lung injury. Transfusion 2014 Apr 25 [Epub ahead of print] 21. Lieberman L, Petraszko T, Yi QL, et al: Transfusion-related lung injury in children: A case series and review of the literature. Transfusion 2014; 54:57–64 22. Lou S, MacLaren G, Best D, et al: Hemolysis in pediatric patients receiving centrifugal-pump extracorporeal membrane oxygenation: Prevalence, risk factors, and outcomes. Crit Care Med 2014; 42:1213–1220 23. Byrnes J, McKamie W, Swearingen C, et al: Hemolysis during cardiac extracorporeal membrane oxygenation: A case-control comparison of roller pumps and centrifugal pumps in a pediatric population. ASAIO J 2011; 57:456–461 November 2014 • Volume 15 • Number 9
Feature Articles 24. Dellinger RP, Levy MM, Rhodes A, et al: Surviving sepsis campaign: International guidelines for management of severe sepsis and septic shock, 2012. Intensive Care Med 2013; 39:165–228 25. Tyree K, Tyree M, DiGeronimo R: Correlation of brain tissue oxygen tension with cerebral near-infrared spectroscopy and mixed venous oxygen saturation during extracorporeal membrane oxygenation. Perfusion 2009; 24:325–331 26. Ang AL, Teo D, Lim CH, et al: Blood transfusion requirements and independent predictors of increased transfusion requirements among adult patients on extracorporeal membrane oxygenation—a single centre experience. Vox Sang 2009; 96:34–43 27. Smith A, Hardison D, Bridges B, et al: Red blood cell transfusion volume and mortality among patients receiving extracorporeal membrane oxygenation. Perfusion 2013; 28:54–60 28. Kumar TK, Zurakowski D, Dalton H, et al: Extracorporeal membrane oxygenation in postcardiotomy patients: Factors influencing outcome. J Thorac Cardiovasc Surg 2010; 140:330–336.e332 29. Burnum JF: Medical vampires. N Engl J Med 1986; 314:1250–1251 30. Corwin HL, Gettinger A, Pearl RG, et al: The CRIT Study: Anemia and blood transfusion in the critically ill–current clinical practice in the United States. Crit Care Med 2004; 32:39–52 31. Yuan S, Tsukahara E, De La Cruz K, et al: How we provide transfusion support for neonatal and pediatric patients on extracorporeal membrane oxygenation. Transfusion 2013; 53:1157–1165 32. Klein HG, Spahn DR, Carson JL: Red blood cell transfusion in clinical practice. Lancet 2007; 370:415–426
Pediatric Critical Care Medicine
33. Luten M, Roerdinkholder-Stoelwinder B, Schaap NP, et al: Survival of red blood cells after transfusion: A comparison between red cells concentrates of different storage periods. Transfusion 2008; 48:1478–1485 34. Bosman GJ, Werre JM, Willekens FL, et al: Erythrocyte ageing in vivo and in vitro: Structural aspects and implications for transfusion. Transfus Med 2008; 18:335–347 35. Tinmouth A, Fergusson D, Yee IC, et al; ABLE Investigators; Canadian Critical Care Trials Group: Clinical consequences of red cell storage in the critically ill. Transfusion 2006; 46:2014–2027 36. Doctor A, Platt R, Sheram ML, et al: Hemoglobin conformation couples erythrocyte S-nitrosothiol content to O2 gradients. Proc Natl Acad Sci U S A 2005; 102:5709–5714 37. Reiter CD, Wang X, Tanus-Santos JE, et al: Cell-free hemoglobin limits nitric oxide bioavailability in sickle-cell disease. Nat Med 2002; 8:1383–1389 38. Marik PE, Sibbald WJ: Effect of stored-blood transfusion on oxygen delivery in patients with sepsis. JAMA 1993; 269:3024–3029 39. Karam O, Tucci M, Bateman ST, et al: Association between length of storage of red blood cell units and outcome of critically ill children: A prospective observational study. Crit Care 2010; 14:R57 40. Gauvin F, Spinella PC, Lacroix J, et al; Canadian Critical Care Trials Group and the Pediatric Acute Lung Injury and Sepsis Investigators (PALISI) Network: Association between length of storage of transfused red blood cells and multiple organ dysfunction syndrome in pediatric intensive care patients. Transfusion 2010; 50:1902–1913
www.pccmjournal.org
813
