Effects of blood transfusion on oxygen extraction ratio and central venous saturation in children after cardiac surgery.

BACKGROUND: Red blood cell transfusion is common in critically ill children after cardiac surgery. Since the threshold for hemoglobin (Hb) transfusion need is not well defined, the threshold Hb level at which dependent critical oxygen uptake-to-delivery (VO2-DO2) status compensation is uncertain.
OBJECTIVES: To assess the effects of blood transfusion on the oxygen extraction ratio (O2ER) and central venous oxygen saturation (ScvO2) to identify a critical O2ER value that could help us determine the critical need for blood transfusion.
DESIGN: Prospective, observational cohort study.
SETTING: Cardiac Surgical Intensive Care Unit at Prince Sultan Cardiac Center in Qassim, Saudi Arabia. PATIENTS AND METHODS: Between January 2013 and December 2015, we included all children with cardiac disease who underwent surgery and needed a blood transfusion. Demographic and laboratory data with physiological parameters before and 1 and 6 hours after transfusion were recorded and O2ER before and 6 hours after transfusion was computed. Cases were divided into two groups based on O2ER: Patients with increased O2ER (O2ER > 40%) and normal patients without increased O2ER (O2ER < =40%) before transfusion. MAIN OUTCOME MEASURE(S): Changes in O2ER and ScvO2 following blood transfusion.
RESULTS: Of 103 patients who had blood transfusion, 75 cases had normal O2ER before transfusion while 28 cases had increased O2ER before transfusion. Following blood transfusion, O2ER and ScvO2 improved in the group that had increased O2ER before transfusion, but not in the group that had normal O2ER before transfusion.
CONCLUSIONS: The clinical and hemodynamic indicators O2ER and ScvO2 may be considered as markers that can indicate a need for blood transfusion. LIMITATIONS: The limitation of this study is the small number of patients that had increased O2ER before transfusion. There were few available variables to assess oxygen consumption.

Red blood cell transfusion is common in critically ill children after cardiac surgery. Improving oxygen transportation and tissue oxygenation are the main goals of blood transfusion in critically ill patients. Thresholds for red blood transfusions in critically ill pediatric cardiac patient are controversial and not well defined. A higher hemoglobin level is frequently targeted to compensate for hypoxia in children with cyanotic heart lesions.1 Some experts advocate transfusion thresholds as high as 14 to18 g/dL in critical cyanotic cases.2 On the other hand, blood transfusion is not without risk. It carries a risk of infection, allergic reaction, morbidity and can lead to transfusion-related acute lung injury.1–3 The critical level of hemoglobin that will improve oxygen transportation and oxygen delivery (DO2) in critically ill children is uncertain. Furthermore, it is unclear at which hemoglobin level blood transfusion has substantial effects on tissue oxygenation or the ability of tissue to extract more oxygen.4 American Heart Association and sepsis campaign guidelines advocate for blood transfusion in critically ill children with hemoglobin less than 10 g/dL and central venous O2 saturation (ScvO2) <70 percent.5

The normal O2ER (oxygen extraction ratio) is around 25% to 30%. When O2ER exceeds 40% to 50%, DO2 starts to be exhausted, leading to tissue dysoxia, a VO2/DO2 dependent status and evidence of O2 debt.6 O2ER and ScvO2 are clinical tools that can help clinicians estimate DO2 and O2 consumption (VO2) status in the body. When hemoglobin drops, the body compensates by multiple physiological compensatory mechanisms, aiming to increase cardiac output and improving DO2 to meet VO2. These compensatory mechanisms are limited and vary within and between individuals and are influenced by different physiological circumstances.7 The threshold Hb level where the compensation process reaches a critical VO2/DO2 status, where VO2 is dependent on DO2, is an important target threshold. There are multiple biomarkers that have been suggested as triggers for transfusion need in critically ill children approaching the critical deflection point of VO2/DO2 dependency status. These may include hemoglobin level, mixed venous oxygen saturation (SVO2), ScvO2 and lactate level. Few adult studies suggested using O2ER as an additional marker or trigger to determine transfusion need in critical condition.8 When O2ER increases in a critically ill patient with low hemoglobin levels, blood transfusion to optimize DO2, the DO2/VO2 balance and O2ER becomes detrimental.5,9,10 Because many cardiac patients have cyanosis and low systemic saturation to start with, the use of O2ER in addition to ScvO2 may reflect better body oxygenation status, VO2/DO2 balance and a critical need for blood transfusion. New monitoring methods such as near-infrared spectroscopy measure regional rather than global perfusion. The role of O2ER as a trigger for blood transfusion in postoperative cardiac children has not been well evaluated. In this study, we assessed the effects of blood transfusion on O2ER and ScvO2 in postoperative cardiac patients. Our goal was to identify a critical O2ER value that can help us better determine the critical need for blood transfusion and avoid an unnecessary and potentially harmful transfusion.

PATIENTS AND METHODS

This prospective non-interventional observational cohort study was undertaken in the Cardiac Surgical Intensive Care Unit (CSICU) at Prince Sultan Cardiac Center, Qassim, Saudi Arabia, between January 2013 and December 2015, after Institutional Review Board approval. We included all children with cyanotic and noncyanotic cardiac disease who underwent corrective or palliative surgery and needed blood transfusion during the postoperative period. No patients were on ECMO during the time of the study. Children with active bleeding (defined as total surgical drainage more than 4 cc/kg /hour) were excluded from study. Patients were not routinely monitored by near-infrared spectroscopy so it was not included in data collection.

No fixed blood transfusion protocol was followed. During the study period, the attending physician decided whether to transfuse blood. The patient’s age, weight, diagnosis of congenital heart disease, RACHS score, and Wernovsky vasoactive score were recorded. The sedation score was not included in the data as some of our patients were extubated at the time of transfusion. Physiological parameters such as heart rate, systolic blood pressure, mean blood pressure, O2 saturation, core temperature, central venous pressure, urine output pre-transfusion and 1 and 6 hours after transfusion were recorded. Laboratory data included complete blood count (CBC), lactate level, arterial blood gas and central venous gas immediately before starting transfusion and then 1 and 6 hours after transfusion. Blood was usually transfused over 2 to 3 hours. No laboratory data were collected during transfusion. The amount of blood transfused was measured as mL/kg. The oxygen extraction ratio was calculated using the formula:

Oxygen extraction before and 6 hours after transfusion were subsequently computed after collecting all data. The posttransfusion values reported in the tables are at 6 hour posttransfusion. Attending physicians were unaware of O2ER values when the decision to transfuse was made. Values of O2ER were calculated at the time of study analysis and not considered upon transfusion because critical O2ER is not well defined, differing from one reference to another.6,16

A value of 50% to 60% O2ER has been suggested as a critical O2ER value.6 Nevertheless, we selected an O2ER of 40% because it is an early and safer warning of tissue O2 debt and m ay reflect better an early prediction of inadequate tissue oxygenation. Cases were divided into two groups based on O2ER results: patients with increased O2ER (O2ER >40%) and without increased O2ER (O2ER ≤40%) before transfusion. The groups were then analyzed using 50% O2ER as a cut point.

Physiological and laboratory data, were recorded. O2ER and SvcO2 values were compared pre- and posttransfusion within each group and between the two groups. Continuous data were analyzed between the two groups using the unpaired t test. Data are presented as mean and standard error or standard deviation where noted and a P value below .05 was considered statistically significant. SSPS version 20 was used for the statistical analysis.

RESULTS

One hundred and three children were transfused after cardiac surgery. Some patients received several blood transfusions and were included in both groups. The mean (SD) Hb level pretransfusion was 9.1 (1.1) g/dL. The average increase in Hb after transfusion was 2.7 (0.03) g/dL (P=.0001). Seventy-five cases had normal O2ER before transfusion with mean (SD) O2ER of 25 (9.6) percent (normal O2ER group) while 28 cases had increased O2ER before transfusion with a mean (SD) O2ER of 46 (4.7) percent (increased O2ER group) (P=.0001).

Mean (SD) Hb levels before transfusion were 9.03 (1.1) and 9.3 (0.1) g/dL for increased and normal O2ER groups, respectively (P=.18) (Table 1). Patients received a mean of 11.0 (0.2) mL/kg of packed red blood cell that led to an average increase in Hb after transfusion of 2.8 g/dL and 2.7 g/dL in the increased and normal O2ER groups, respectively.

Table 1

Demographic data in patients grouped by oxygen extraction ratio status before transfusion.

Variables	Increased O2ER (>40%) before transfusion (n=28)	Normal O2ER (≤40%) before transfusion (n=75)	P
Age (month)	10.8 (3.0)	14.3 (2.8)	.47
Weight (kg)	6.23 (0.5)	6.7 (0.6)	.63
Hemoglobin (g/dL)	9.03 (1.1)	9.3 (1.1)	.18
Cyanotic/non-cyanotic	8/20 (40%)	33/42 (78%)	.18
RACHS-1 score	2.5 (0.2)	2.7 (0.07)	.13

Values are mean (standard deviation) or n (%). P values from unpaired t test.

Patients from both groups had a significant improvement in Hb level following transfusion. These changes were statistically significant between pre- and posttransfusion Hb level (Tables 2 and 3), but there were no statistical differences in Hb level between the increased and normal O2ER groups during transfusion (Table 4).

Table 2

Patients with normal oxygen extraction ratio(≤40%) before transfusion (n=75).

Variables	Pretransfusion	Posttransfusion	P
Heart rate (bpm)	138.2 (2.2)	134 (2.3)	.18
Systolic blood pressure (mm Hg)	85 (1.4)	87.2 (1.4)	.26
Mean blood pressure (mm Hg)	65.2 (1.2)	66.7 (1.2)	.37
Diastolic blood pressure (mm Hg)	48.37 (1.2)	49.98 (1)	.3
O2 arterial saturation (%)	91.7 (0.9)	92.5 (0.9)	.5
Central venous pressure (mm Hg)	9.6 (0.36)	10 (0.37)	.44
Urine output(mL/kg)	4.3 (0.3)	4.5 (0.2)	.58
Core temperature oC	36.4 (0.1)	36.5 (0.1)	.48
Hemoglobin(g/dL)	9.3 (1.1)*	12 (1.5)*	.0001
SvcO2 (%)	67.5 (1.4)	67.5 (1.2)	1
Lactic acid (mmol/L)	1.9 (0.2)	1.6 (0.2)	.3
SvcCO2 (%)	5.6 (0.4)	6.02 (0.4)	.45
Vasoactive score	7.5 (0.5)	7.6 (0.5)	.88
O2ER	25.0 (9.6)*	25.8 (9.8)*	.6

Central venous oxygen saturation (ScvO2); Oxygen extraction ratio (O2ER). P values from unpaired t test.

Standard deviation.

Table 3

Patients with increased oxygen extraction ratio (>40%) before transfusion (n=28).

Variables	Pretransfusion	Posttransfusion	P
Heart rate (bpm)	138 (3.8)	135.7 (3.6)	.56
Systolic blood pressure (mm Hg)	84.17 (1.8)	84.35 (1.7)	.9
Mean blood pressure (mm Hg)	66.17 (1.5)	64.92 (1.9)	.6
O2 arterial saturation (%)	89.5 (1.6)	89.66 (1.5)	.99
Central venous pressure (mm Hg)	10.14 (0.4)	10.25 (0.38)	.84
Urine output (mL/kg)	3.7 (0.27)	4.2 (0.29)	.21
Core temperature oC	36.45 (0.16)	36.45 (0.11)	.99
Hemoglobin (g/dL)	9.03 (1.1)*	11.8 (1.4)*	.0001
SvcO2 (%)	51.3 (1.8)	58.9 (2)	.0065
Lactic acid (mmol/L)	1.7 (0.2)	1.4 (0.1)	.18
SvcCO2 (%)	8.47 (0.82)	6.6 (0.6)	.07
Vasoactive score	7.9 (0.5)	8 (0.5)	.88
O2ER	11.8 (1.4)*	11.8 (1.4)*	.0001

Central venous oxygen saturation (ScvO2); Oxygen extraction ratio (O2ER). P values from unpaired t test.

Standard deviation.

Table 4

Physiologic and laboratory data in patients grouped by oxygen extraction ratio status before transfusion.

Variables	Increased O2 ER (>40%) before transfusion (n=28)	Normal O2 ER (≤40%) before transfusion (n=75)	P
Pretransfusion heart rate (bpm)	138 (3.8)	138 (2.2)	.99
Systolic blood pressure (mm Hg)	84.17 (1.8)	85 (1.4)	.74
Mean blood pressure (mm Hg)	66.17 (1.5)	65.2 (1.2)	.65
Change in mean blood pressure (mm Hg)	0.02 (0.3)	0.36 (0)	.06
Arterial saturation (%)	89.5 (1.6)	91.7 (0.9)	.21
Core temperature oC	36.45 (0.16)	36.4 (0.1)	.79
Urine output (mL/kg)	3.7 (0.27)	4.3 (0.3)	.25
Change in hemoglobin (g/dL)	2.7 (0.06)	2.7 (0.1)	.99
Vasoactive score	7.9 (0.5)	7.5 (0.5)	.65
Lactic acid (mmol/l)	1.7 (0.2)	1.9 (0.2)	.57
ScvO2 (%)	51.3 (1.8)	67.5 (1.4)	.0001
Change in ScvO2 (%)	7.6 (0.2)	0 (0.2)	.0001
ScvCO2 (%)	8.47 (0.82)	5.6 (0.4)	.0006
O2ER	46 (0.9)	25 (1.1)	.0001
Change in O2ER	9.6 (0.9)	0.8 (0)	.0001

Central venous oxygen saturation (ScvO2); Oxygen extraction ratio (O2ER). P values from unpaired t test.

The increased O2ER group had a lower ScvO2 saturation in comparison to the normal O2ER group, 51.3 (1.8) percent versus 67.5 (1.4) percent (P=.0001). Following blood transfusion mean (SD) O2ER improved from 46 (4.7) percent to 36.4 (9.5) percent (P=.0001), in the increased O2ER group while in the normal group there were no significant changes in O2ER (25 [9.6] percent to 25.8 [9.8] percent, P=.6). Increments of central venous saturation change (ΔScvO2) after transfusion were statistically higher in the increased O2ER group versus the normal O2ER group (P=.0001). There were no significant differences in heart rate, systolic blood pressure or mean blood pressure before or after transfusion in either group. There were no statistically significant differences in lactate level.

In a subgroup analysis, in patients with hemoglobin >10 g/dL and increased O2 extraction there was a significant improvement in the change in ScvO2 (increment of ScvO2 posttransfusion) and O2ER after transfusion in comparison with normal ScvO2 and O2ER values in the same subgroup (Table 5). Likewise, in patients with hemoglobin <10 g/dL there were significant improvements in change in ScvO2 and O2ER after transfusion in comparison with the normal group (Table 5). In a subgroup analysis of patients with O2ER >50%, O2ER improved in the increased versus the normal O2ER group (Table 6).

Table 5

Changes in oxygen extraction ratio and central venous saturation after blood transfusion in patients with HB ≥10 or <10 in patients grouped by increased and normal oxygen extraction ratio.

Variable	Low HBHB <10 (n=75)		P	95% CI	High HbHB ≥10 (n=28)		P	CI
	Increased group Pretransfusion O2ER >40% (n=20)	Normal group Pretransfusion O2ER <40% (n=55)			Increased group Pretransfusion O2ER >40% (n=8)	Normal group Pretransfusion O2ER <40% (n=20)
Change in ScvO2 Posttransfusion	7.3 (0.4)	0.6 (0.1)	.0001	(6.1,7.3)	9.2 (0.56)	2.8 (0.7)	.0001	(3.9,8.8)
Change in O2ER Posttransfusion	8.5 (0.9)	1.4 (0)	.0001	(6.03,8.16)	13.07 (2.6)	0.3 (0.1)	.0001	(9.6,15.9)

Central venous oxygen saturation (ScvO2); Oxygen extraction ratio (O2ER). P values from unpaired t test for changes in values from the pretransfusion state.

Table 6

Comparison of patients with increased (n=6)and normal oxygen extraction ratio (n=97) by oxygen extraction ratio greater than 50%.

	Increased O2ER	Normal O2ER	P
Pretransfusion
Heart rate (bpm)	137.4 (1.9)	147.3 (7.9)	.13
Mean blood pressure (mm Hg)	59.16 (3.6)	66 (0.9)	.09
Urine output (mL/kg)	3.5 (0.36)	4.48 (0.19)	.2
Hemoglobin (g/dL)	8.9 (0.49)	9.6 (0.29)	.55
Lactic acid (mmol/l)	2.13 (0.6)	1.8 (0.17)	.63
ScvO2 (%)	44.7 (2.2)	64.22 (1.3)	.0004
O2ER (%)	53 (0.6) (98)	29.31 (1.19) (97)	.0026
Posttransfusion O2ER (%)	43.0 (2.9)	27.7 (1.0)	.0001

Central venous oxygen saturation (ScvO2); Oxygen extraction ratio (O2ER). P values from unpaired t test.

DISCUSSION

Transfusion of red blood cells during and after cardiac surgery is a common practice. When the degree of anemia compromises oxygen delivery and leads to organ hypoxia, the need for blood cell transfusion becomes vital.4 Although the evidence is poor, most RBC transfusions are given because the hemoglobin concentration has fallen below a static predefined threshold that is perceived to be associated with insufficient arterial oxygen, but this hemoglobin concentration varies between and even within institutions. Hemoglobin level is not sufficient as the only trigger for blood transfusion. The hemoglobin level may vary not only on the basis of cardiac lesion and systemic saturation, but also on the basis of cardiac function and adaption of the circulation to changes in loading conditions imposed by the operative correction. These factors may not only change from heart lesion to lesion, but on the hearts adaption to changes in cardiac loading conditions. Lots of studies have examined other potential triggers to the decision to transfuse blood. In TRIPICU study11 the authors showed clearly that blood transfusion did not improve the outcome in a restricted or liberal group even in Fontan and Glenn patients. Goal-directed therapy is probably the solution for this dilemma. From the clinical perspective, it is desirable to find more reliable and measurable predictors for adequate organ oxygenation and to replace the static hemoglobin transfusion trigger with a physiologic trigger. Mixed venous oxygen saturation (MVO2) sampled from the pulmonary artery is ideal for measurements of DO2 and VO2. Because of the difficulty in obtaining MVO2 in critically ill children, ScvO2 was suggested as a surrogate for MVO2.12 In clinical practice ScvO2 taken from the superior vena cava may be the most practical physiological marker of oxygen demand because it reflects the DO2/VO2 balance.4

The targeted recommended ScvO2 is 70% in critical children with a normal structural heart,5 but in children with cardiac disease and cyanotic heart disease the critical ScvO2 is not clearly defined. Interpreting and defining a critical ScvO2 value in children with cyanotic heart lesion is more difficult since many of them have arterial O2 saturation (SaO2) ranging between 75% and 85%. Examining O2ER in those patients is of interest and has clinical relevance. In 1982 Jamieson13 found that MVO2 saturation changes take place before any changes in mean arterial pressure or heart rate and it is well correlated with the cardiac index. Staphen et al12 found that increased oxygen extraction is associated with increased length of stay and multiorgan failure in the ICU. To what extent an increase in hemoglobin concentration will enhance tissue and organ oxygenation is unclear and still under investigation. Herbet et al14 reviewed 18 clinical studies investigating the effects of RBC transfusions on DO2/VO2 relationship in critically ill patients. They reported that improved VO2 was documented in only five studies. O’Farrell conducted a pilot study in 2006 that measured the relationship between O2ER and postoperative RBC transfusions in cardiac surgery and concluded that elevated O2ER may be a more appropriate transfusion trigger than low hemoglobin concentration and its use may reduce inappropriate transfusions.9 David et al10 found a significant improvement in O2ER values at 15 and 120 minutes after transfusion and suggested including O2ER into transfusion decision triggers assuming that it would reduce postcardiac surgery RBC transfusions.10 On the other hand, in 2014 Fiser et al17 investigated the effect of RBC transfusion in 45 pediatric patients with ECMO and found that transfusion did not significantly alter global tissue oxygenation. In 2013 Mungayi et al8 studied the direct effect of blood transfusion on oxygen extraction in a general adult intensive care unit and found no statistically significant difference in oxygen extraction pre- and posttransfusion. In these two previous studies, most transfusions were given when the patient did not appear to be oxygen delivery dependent. Thus, Mungayi et al recommended other studies to investigate the effects of blood transfusion in increasing pretransfusion O2 extraction. Whether RBC transfusion may increase DO2 and the VO2/DO2 balance depends on how severely tissue oxygenation is increased.7 In critically ill patients suffering from acute respiratory syndrome with oxygen supply dependency status and metabolic acidosis, Kruse et al15 demonstrated that blood transfusions contributed to the augmentation of DO2 thereby providing a significant increase in VO2. Marino16 pointed out that an O2ER more than 50% is an indirect marker of tissue dysoxia or impending dysoxia and suggested that this number could be used as an indicator for transfusion trigger.

We found statistically significant differences between O2ER and ScvO2 pre- and posttransfusion in the increased O2ER group (both 40% or 50% O2ER) but not the normal O2ER group. Both groups had similar in hemoglobin levels pretransfusion, but there may have been differences in cardiac output but still increase O2ER had benefit from blood transfusion even if increase O2ER is not because of low hemoglobin. These changes were observed even when hemoglobin was more than 10 g/dL when O2ER was increased, which suggested that hemoglobin concentration is not enough as a sole marker to guide blood transfusion. Furthermore, we observed no changes in urine output or lactate level, which is consistent with the opinion of Jamieson13 who stated that decreased MVO2 and increased O2ER are early signs of dysoxia which can be picked up earlier than hypotension and acidemia. We found that neither a hemoglobin value nor clinical assessment alone was enough to decide whether to do a blood transfusion. One advantage of using O2ER in guiding transfusion is that in patients with a cyanotic heart lesion as it reflects ratio rather than absolute number where many patients may have systemic O2 saturation ranging between 70%–85% and ScvO2 below 70%.

The unique value of this study is the reliance on measures of oxygen delivery independent of actual hemoglobin level. However, the limitation of this study is the small number of patients in the increased O2ER group and the lack of few variables to assess oxygen consumption rather than core temperature. Large randomized prospective trials are needed to support our findings and to establish a goal-directed blood transfusion protocol. Furthermore, the decision to undertake a blood transfusion is a clinical decision that incorporates many elements such as patient stability, level of hemoglobin, patient preference, the presence of a cyanotic heart lesion and many others. Postoperatively, in children undergoing cardiac surgery with an O2ER more than 40%, blood transfusion improves O2ER and ScvO2. O2ER and ScvO2 might be useful as additional blood transfusion triggers.