Association of haematocrit and red blood cell transfusion with outcomes in infants with shunt-dependent pulmonary blood flow and univentricular physiology

Abstract

Background

The aim of this study was to investigate the association between red blood cell (RBC) transfusion and haematocrit values with outcomes in infants with univentricular physiology undergoing surgery for a modified Blalock-Taussig shunt.

Material and methods

This study included infants ≤ 2 months of age who underwent modified Blalock-Taussig shunt surgery at the Arkansas Children’s Hospital (2006–2012). Infants undergoing a Norwood operation or Damus-Kaye-Stansel operation with modified Blalock-Taussig shunt were excluded. Demographics, pre-operative, operative, daily laboratory data, and post-operative variables were collected. We studied the association between haematocrit and blood transfusion with a composite clinical outcome. Multivariable logistic regression models were fitted to study the probability of study outcomes as a function of haematocrit values and RBC transfusions after operation.

Results

Seventy-three patients qualified for inclusion. All study patients received blood transfusion within the first 48 hours after heart surgery. The median haematocrit was 44.3 (interquartile range [IQR] 42.5–46.2), and the median volume of RBC transfused was 28 mL/kg (IQR, 10–125) in the first 14 days after surgery. The overall in-hospital mortality rate was 13.6% (10 patients). A multivariable analysis adjusted for risk factors, including weight, prematurity, cardiopulmonary bypass and postoperative need for nitric oxide and dialysis, revealed no association between haematocrit values and RBC transfusion with the composite clinical outcome.

Discussion

We did not find an association between higher haematocrit values and increasing RBC transfusions with improved outcomes in infants with shunt-dependent pulmonary blood flow and univentricular physiology. The power of our study was small, which prevents any strong statement on this lack of association. Future multi-centre, randomised controlled trials are needed to investigate this topic in further detail.

Introduction

Red blood cell (RBC) transfusion strategies have been a topic of much discussion, particularly in the realm of paediatric cardiac surgery. In both adult and paediatric literature, RBC transfusions have been shown to be independently associated with increased morbidity and mortality^1–3. Some of the complications associated with RBC transfusion include infections, electrolyte abnormalities, immunomodulation, allergic reactions, transfusion-related acute lung injury and circulatory overload⁴. Despite these complications and their risks, liberal blood transfusion protocols for paediatric cardiac surgery continue to persist. Certain authors have proposed that a higher haematocrit may be beneficial for cyanotic patients in an effort to improve oxygen delivery^5–6. Neonatal age, cyanotic heart disease, and paediatric cardiopulmonary bypass circuits are factors known to increase the need for transfusion among children undergoing cardiac surgery^7–10.

Existing literature describing outcomes of paediatric patients undergoing cardiac surgery receiving blood transfusion have included heterogeneous populations of patients undergoing operations of varying complexity^11–13. Despite advances in post-operative management of critically ill children with heart disease, the effects of blood transfusion on conventional outcome measures in a homogenous population of children undergoing cardiac surgery are unclear. As blood transfusion is a modifiable factor in post-operative care of children undergoing heart surgery, we sought to explore the association of haematocrit and blood transfusion with outcomes in infants with univentricular physiology undergoing placement of a modified Blalock-Taussig shunt (mBTS). We hypothesised that higher haematocrit values and increased blood transfusions in the first 2 weeks after the mBTS placement are associated with improved outcomes.

Materials and methods

We performed a single-centre, retrospective, observational study in a 15-bed paediatric cardiovascular intensive care unit (ICU) in a tertiary-care, academic children’s hospital during the period from January 2006 to December 2012. The study included all infants ≤2 months of age in whom a mBTS was placed at the Arkansas Children’s Hospital. Subjects excluded from the study were infants undergoing a Norwood operation or Damus-Kaye-Stansel operation, infants operated upon in other institutions and subsequently transferred to our institution, infants receiving extracorporeal membrane oxygenation in the post-operative period, infants undergoing orthotopic heart transplantation or receiving a ventricular assist device during the same hospital stay, and infants with a “Do Not Resuscitate” order. We identified the potential subjects by querying the departmental surgical database and the hospital medical records. The Institutional Review Board of the University of Arkansas Medical Sciences approved the study and the need for informed consent was waived.

Demographics, pre-operative, operative, laboratory data, and post-operative variables were collected. The demographic information collected included age, weight, gender, underlying genetic abnormality, prematurity, and associated diagnoses, as well as the presence of an antenatal diagnosis. Intra-operative variables collected included cardiopulmonary bypass (CPB) time, cross clamp time, and need for an open chest after heart surgery. The laboratory indices collected included daily haematocrit, creatinine and lactate values for the first 14 days after the operation. We also recorded the volume of RBC transfusions in millilitres per kilogram (mL/kg) during the operation and daily for the first 14 days after it. At our centre there is currently no protocol in place for RBC transfusions in children undergoing heart operations for congenital heart disease. However, the most common indications for RBC transfusion in these patients include: (i) haematocrit <30%, (ii) continued bleeding, and/or (iii) evidence of decreased oxygen-carrying capacity, for example, demonstrated by a low mixed venous saturation and/or tachycardia. We consider post-operative bleeding to be significant if blood loss exceeds 10 mL/kg/hour or 4 mL/kg/hour for three consecutive hours. The other post-operative variables collected in our study included the use of nitric oxide, dialysis after the heart operation and positive blood cultures.

The primary outcome measures evaluated included mortality, length of stay in the ICU, duration of mechanical ventilation, and days to chest closure. The secondary outcome measures evaluated included lactate levels, estimated glomerular filtration rate (eGFR), and inotrope score in the first 14 days after heart surgery. Both unadjusted and adjusted outcomes were evaluated. In addition we also defined a composite clinical outcome, taking into account primary and secondary outcome measures. The composite clinical outcome variable was defined as in-hospital death or any one of the following morbid events: prolonged length of stay in the ICU, prolonged duration of mechanical ventilation, prolonged duration of an open chest, elevated lactate levels, elevated inotrope score, and elevated eGFR. An open chest was defined as a sternum not closed in a patient who arrived in the cardiovascular ICU. The inotrope score was calculated using the following equation: Dosages of dopamine + dobutamine (in μg/kg/min) + (Dosages of adrenaline + noradrenaline + isoproterenol [in μg/kg/min]) ×100 + dosages of milrinone (in μg/kg/min) ×15¹⁴. eGFR was assessed using the modified Schwartz formula (0.413×height [cm]/serum creatinine [mg/dL])¹⁴. Length of ICU stay was calculated from the date of post-operative admission until discharge from the ICU. The duration of mechanical ventilation was from the date of postoperative admission until extubation. The 75^th percentiles of the cohort data distribution were used to establish cut-off values for defining prolonged length of stay in the ICU, prolonged duration of mechanical ventilation, prolonged duration of an open chest, elevated lactate level, elevated inotrope score, and elevated eGFR. This is the same approach as used by Gaies et al.¹⁵ and Wernovsky et al.¹⁶ in published literature.

Statistical analysis

Continuous variables are presented as the median and interquartile range (IQR) between Q1 and Q3, where Q1 is the 25^th percentile, and Q3 is the 75^th percentile, whereas categorical variables are presented as numbers and percentages. p-values were calculated using the chi-square test and/or Fisher’s exact test of independence for categorical variables and the Wilcoxon rank-sum test for continuous variables. A power analysis demonstrated that a sample size of 80 patients was necessary to yield at least 80% power to detect a difference of 10% in mortality rates between patients with mean blood transfusion levels and patients with blood transfusion one standard deviation above the mean. This is equivalent to detecting an odds ratio in mortality of 2.67 between the two blood transfusion levels. The above calculation was done assuming blood transfusion levels are normally distributed and a two-sided significance level of 5%.

Cox proportional hazards models were fitted to study the probability of study outcomes as a function of haematocrit values and RBC transfusions after operation. Variables with a p value of ≤0.1 in the univariate analysis were entered into the multiple regression models. Variables with ≥20% missing values were not considered for inclusion into the multivariate models. The following variables were selected for multivariable models: weight, prematurity, cardiopulmonary bypass, and postoperative need for nitric oxide and dialysis. The models’ results are expressed in terms of adjusted odds ratios for categorical variables and hazard ratios for continuous variables, 95% confidence interval, and p values. Several additional multiple logistic analyses were performed to explore variables left out of the model and to achieve a parsimonious model. A model’s goodness-of-fit was evaluated using the Hosmer-Lemeshow test, and the discrimination of the model was assessed using the area under the receiver operating characteristic (ROC) curve. Analyses were performed using STATA/MP, version 11.1 software (Stata^® Corp LP, College Station, TX, USA).

Results

During the study period, 73 patients qualified for inclusion in the study. All study patients received blood transfusion within the first 48 hours after their heart operation. The median age of the patients was 9 days (IQR: 6, 20) and their median weight was 3.2 kg (IQR: 2.8, 3.6). Of the study patients, 48 (67%) patients were male, 5 (6%) patients had an underlying genetic abnormality, and 17 (6%) patients were classified as premature (≤36 weeks gestational age) (Table I). Sixteen (21%) patients were admitted to the ICU with an open chest. Nine (12%) patients required dialysis after surgery, and 11 (15%) patients had positive blood cultures during their post-operative course. Ten patients died in hospital, for an overall in-hospital mortality rate of 13.6%.

Table I.

Patients’ characteristics and outcomes.

Patients’ variables	N (%) or median (25th, 75th quartiles) (N=73)
Age (days)	9 (6, 20)
Weight (kg)	3.1 (2.8, 3.6)
Male gender	48 (66%)
Underlying genetic abnormality	7 (10%)
Prematurity (≤36 weeks)	17 (23%)
Antenatal diagnosis	18 (25%)
CPB time (minutes)	55 (47, 91)
Cross clamp time (minutes)	10 (6, 12)
Delayed sternal closure	16 (22%)
Use of nitric oxide	13 (18%)
Duration of nitric oxide (days)	5 (3, 9)
Dialysis	9 (12%)
Positive blood cultures	11 (15%)
Use of factor VII	6 (8%)
Reoperation for bleeding	4 (5%)
Mortality	10 (13%)
Duration of open chest (days)	7.5 (2, 16)
Duration of mechanical ventilation (days)	5 (2, 14)
Length of ICU stay (days)	25 (13, 44)

CPB: cardiopulmonary bypass; ICU: Intensive Care Unit.

The following patients were included in our study: 33% (24/73) with tetrology of Fallot and pulmonary atresia; 16% (12/73) with pulmonary atresia and intact ventricular septum; 16% (12/73) with double outlet right ventricle and mitral atresia; 14% (10/73) with malaligned complete atrioventricular canal and hypoplastic (right or left) ventricle; 11% (8/73) with tricuspid valve atresia (with normal or transposed great arteries); 5% (4/73) with Ebstein’s anomaly; and 4% (3/73) with double inlet right ventricle (with normally related or transposed great arteries).

The median RBC transfusion was 10 mL/kg (IQR: 0, 33) in the first 48 hours after heart surgery, 23 mL/kg (0, 82) in the first 7 days after heart surgery, and 28 mL/kg (10, 125) in the first 14 days after heart surgery. The median haematocrit was 45.2 (IQR: 42.5, 48.1) in the first 48 hours after the heart operation, 44.3 (IQR: 42.4, 46.2) in the first 7 days after the heart operation, and 44.3 (IQR: 42.5, 46.2) in the first 14 days after the heart operation.

The median length of stay in the ICU was 25 days (IQR: 13, 14), the median duration of mechanical ventilation was 5 days (IQR: 2, 14) and the median time to chest closure was 7.5 days (IQR: 2, 16) for the patients who had an open chest after their heart operation. In the 14 days after heart surgery, the median eGFR was 49 mL/min (37, 60), the median lactate value was 1.6 mmol/L (1.3, 2.1), and median inotrope score was 2.1 (0.8, 6.3).

Adjusted outcomes are documented in Tables II and III. A multivariable analysis adjusted for risk factors, including weight, prematurity, cardiopulmonary bypass and post-operative need for nitric oxide and dialysis, revealed no association between haematocrit and RBC transfusion and the composite clinical outcome. Adjusted primary and secondary outcomes are presented in e-Tables I and II in the online supplementary content. Haematocrit and RBC transfusion were not associated with mortality, length of ICU stay, duration of mechanical ventilation, and days to chest closure. After adjusting for risk factors, inotrope score and eGFR in the first 14 days after heart surgery were not associated with haematocrit and RBC transfusion. However, lactate levels were significantly elevated for the first 7 days after heart surgery with increased RBC transfusions. Study models had good performance and fit (e-Table III of the online supplementary content).

Table II.

Study outcomes based on post-operative haematocrit.

Outcome	Mean haematocrit		Lowest haematocrit		Highest haematocrit
	OR (95% CI)	p-value	OR (95% CI)	p-value	OR (95% CI)	p-value
Composite clinical outcome	1.06 (0.88–1.28)	0.54	1.10 (0.94–1.28)	0.26	1.00 (0.90–1.11)	0.98
Mortality	1.22 (0.81–1.82)	0.35	1.13 (0.76–1.68)	0.54	1.17 (0.94–1.47)	0.17
Prolonged length of ICU stay	1.27 (1.00–1.62)	0.07	1.12 (0.92–1.36)	0.26	1.08 (0.95–1.22)	0.24
Prolonged duration of MV	1.11 (0.85–1.44)	0.46	1.07 (0.86–1.35)	0.54	0.99 (0.83–1.18)	0.90
Prolonged duration of open chest	0.80 (0.56–1.17)	0.25	0.83 (0.60–1.14)	0.24	1.05 (0.84–1.32)	0.67
Elevated lactate	0.92 (0.75–1.14)	0.44	1.02 (0.84–1.23)	0.87	0.90 (0.77–1.05)	0.17
Elevated inotrope score	1.22 (0.91–1.65)	0.19	1.20 (0.91–1.58)	0.20	1.05 (0.87–1.25)	0.62
Elevated eGFR	0.97 (0.78–1.20)	0.76	0.96 (0.80–1.14)	0.63	0.98 (0.86–1.13)	0.81

OR: odds ratio; CI: confidence interval; ICU: Intensive Care Unit; MV: mechanical ventilation; eGFR: estimated glomerular filtration rate.

The adjusted multivariable model includes weight, prematurity, cardiopulmonary bypass, and post-operative need for nitric oxide and dialysis.

Table III.

Study outcomes based on blood transfusion after cardiac surgery.

Outcome	OR (95% CI)	p-value
Composite clinical outcome	1.23 (0.99–1.53)	0.06
Mortality	0.98 (0.89–1.08)	0.68
Prolonged length of ICU stay	1.02 (0.95–1.08)	0.60
Prolonged duration of MV	1.00 (0.92–1.08)	0.93
Prolonged duration of open chest	1.03 (0.95–1.11)	0.50
Elevated lactate	1.06 (0.97–1.15)	0.22
Elevated inotrope score	1.11 (0.96–1.29)	0.17
Elevated eGFR	0.98 (0.90–1.06)	0.60

CI: confidence interval; ICU: Intensive Care Unit; MV: mechanical ventilation; eGFR: estimated glomerular filtration rate.

The adjusted multivariable model includes weight, prematurity, cardiopulmonary bypass, and post-operative need for nitric oxide and dialysis.

Discussion

Despite their uniquely compromised physiology, children with univentricular physiology undergoing surgery for placement of a mBTS did not seem to benefit from increasing RBC transfusions and higher haematocrit values in the immediate post-operative period. All patients in our cohort required blood transfusions after heart surgery. Increasing RBC transfusions or higher haematocrit values were not associated with improved outcomes, such as lower mortality, a shorter stay in the ICU, a shorter period of conventional mechanical ventilation, or earlier chest closure. Our study also did not demonstrate any association between increasing RBC transfusions and higher haematocrit values with risk adjusted outcomes for lactate levels, inotrope score, and eGFR in the first 14 days after the heart operation.

RBC transfusion has been independently associated with worsened outcomes both in adult and paediatric patients^2,17. In a study of 60 patients with single-ventricle physiology undergoing cavo-pulmonary connection, children randomised to a liberal transfusion strategy (haemoglobin ≥13.0 g/dL) did not benefit in immediate outcome measures compared to those managed with a restrictive transfusion strategy (haemoglobin <9.0 g/dL)¹⁸. However, this study was inadequately powered to investigate clinical outcomes such as mortality or length of ICU stay¹⁸. In another study of 94 paediatric patients undergoing heart transplantation, increasing amounts of RBC transfusions were associated with longer time spent in the ICU, higher inotrope scores, and major adverse events¹⁹. Our results did not show an association between RBC transfusion and primary or secondary study outcomes, suggesting the increased cost and potential risks of administering RBC transfusions may not result in better outcomes for patients.

It has been hypothesised that higher haemoglobin levels are advantageous in the clinical management of cyanotic patients^5,6. Oxygen delivery is directly related to cardiac output, oxygen saturation, haemoglobin concentration and dissolved oxygen in the blood. In an effort to improve oxygen delivery in patients with cyanotic heart disease it has been theorised that higher haemoglobin concentration may be desirable. However, transfused blood may not deliver oxygen in the most efficient manner; this possibly may be due to decreased RBC deformability and RBC-dependent vasoregulatory function and an increased haemoglobin affinity for oxygen^20,21. Our results showed that higher haematocrit levels and increased amounts of RBC transfusions did not give rise to significantly higher eGFR or lower lactate values, both of which correlate with improved oxygen delivery. In fact, increasing amounts of blood transfusion were associated with significantly elevated lactate levels in the first 7 days after heart surgery in our study. Our study did not demonstrate an association between the amount of transfused blood and inotrope scores. The inotrope score can be described as a numerical assessment of pharmacological cardiovascular support. In a recent study, it was demonstrated that increasing RBC transfusions were associated with higher inotrope scores in the first 24 hours after transplantation¹⁹. However, our study failed to demonstrate this relationship.

Our study has several limitations. The results generated by this single-centre study may be unique to our institution and may not be generalizable to all centres. The retrospective nature of the study renders it susceptible to study design flaws and bias. The number of patients in the study was small, which limits our capability to determine precise associations of RBC transfusion and haematocrit values with in-hospital outcomes. All confounding risk factors shown in previous studies to be associated with worsening outcomes were included in our multivariable analysis; however, there may be other variables affecting the outcome that were not included. It is possible that physiological perturbations that may be seen during blood loss can also serve as potential risk factors in affecting outcomes. Our institution lacks a transfusion protocol. The decision to transfuse blood is based on clinical grounds and the attending clinician’s preference. In the absence of a protocol, we cannot completely exclude any bias that may have led to excessive transfusions in these patients. We also lacked data on transfusion of other blood products such as platelets, fresh-frozen plasma, and cryoprecipitate. The number of different blood donors to whom our study patients were exposed is also missing from our study. We could not find any association between amount of blood transfused and haematocrit values as we lacked data on the amount of blood loss. We also lacked haemodynamic data (such as central venous pressures, heart rate, blood pressures, cerebral oximetry) as well as information on the chest tube output. However, we did have data on indirect markers of bleeding such as factor VII use as well as the need for reoperation for bleeding. Our unit has no well-defined protocol for ICU and hospital discharge. In the absence of such a protocol we may have over-estimated or under-estimated the length of stay in the ICU and in hospital. There was a lack of information on the indication for nitric oxide, Qp/Qs and hypoxia. Although to our knowledge, there was no change in our clinical practice during the study period, it is possible that there were some unrecognised changes that could not have been teased out due to the retrospective nature of this study. Given all these shortcomings, we recommend that the findings of this study are rigorously evaluated in future, prospective, multicentre trials. This study has prompted our institution to develop a transfusion protocol for children undergoing heart surgery for congenital heart disease.

Conclusion

We did not find an association between higher haematocrit values and increasing RBC transfusions and improved outcomes in infants with shunt-dependent pulmonary blood flow and univentricular physiology. The power of our study was, however, too small to enable any strong statements to be made. Future multicentre, randomised controlled trials are needed to investigate this topic in further detail.