Impact of restrictive versus liberal transfusion and clinical outcomes in critically ill children: A retrospective observational study

Abstract Background and Aims Critically ill children with anemia often requires blood transfusion, which can cause several complications. It is important to decide when to start the red blood cell (RBC) transfusion; however, the guidelines is still lacking. The aim of this study was to compare restrictive and liberal transfusion strategy. Methods This is an observational retrospective study of critically‐ill children who receive RBC transfusion. Subjects categorized into two groups by initial hemoglobin (Hb), that is, restrictive (Hb ≤ 7 g/dl) and liberal (Hb ≤ 9.5 g/dl) strategy. In each group, subjects categorized based on: (1) Hb increment: high (increased ≥2.5 g/dl) and low (increase <2.5 g/dl) and (2) final Hb level: low (<7.0 mg/dl), moderate (7.0–10.0 mg/dl), and high (>10.0 mg/dl). Patient with hematologic or congenital disorder, severe malnutrition, chronic infection‐related anemia, and transfusion in Hb level ≥9.5 g/dl were exclude. Each patients were evaluated for the clinical outcome, which is: intensive care length of stay (IC‐LOS), length of mechanical ventilation (LoMV), and mortality rate. Results Clinical outcome and mortality rates of both transfusion strategies are similar. The mortality rates were lower in higher Hb increment and final Hb level (p = 0.04 and p = 0.01, respectively). Multivariate analysis in all groups revealed mortality rate had moderate correlation with Hb increment (odds ratio [OR] = 0.694, 95% confidence interval [CI] 0.549–0.878; p = 0.002) and moderate correlation (OR = 0.642, 95% CI 0.519–0.795; p = 0.000) with final Hb level. The similar results was found after categorization based on transfusion strategy. Conclusion We conclude the restrictive and liberal transfusion strategy have a similar effect to IC‐LOS, LoMV, and mortality rate. High Hb increment (≥2.5 g/dl) and moderate‐high final Hb (≥7.0 g/dl) after transfusion reduce the mortality rate.


| INTRODUCTION
Critical illness in children often results in anemia due to various mechanism, such as decrease of erythropoietin activity, excess usage of iron, or bleeding. 1 Anemia could aggravate the patient's condition by reducing tissue oxygen delivery. Anemia incidence in children with critical illness who require pediatric intensive care unit (PICU) admission reaches 41% in 1-3 days, and around 74% of them required red blood cell (RBC) transfusion. 2 Packed red cell (PRC) transfusion provides addition of hemoglobin (Hb) to increase blood oxygen capacity and restoration of intravascular volume to maintain cardiac output and organ perfusion. 3 Massive transfusion might be demanded in certain situation of acute blood loss, but this treatment also brings some sort of complication. Acute blood loss causing acute anemia will induce platelet function disruption which ends on loss coagulopathy.
Massive transfusion causing dilutional coagulopathy by activating extrinsic coagulation process and disrupt the hemostatic mechanism even further. 4 The other harm caused is citrate overload that can significantly induce hypocalcemia and metabolic acidosis. This condition also contributes to alteration of hemostatic process. 5 Another complication of transfusion also including transfusionrelated acute lung injury, allergic reaction and transfusionassociated circulatory overload. 6 Thus, the transfusion treatment must be decided carefully to achieve the purpose and minimize the complication.
Recently, none of the guideline fully describes the strategy of blood transfusion for children with critical illness. Restrictive transfusion could reduce the frequency of transfusion, which might be beneficial than liberal transfusion strategy. TRIPICU study compared restrictive transfusion (Hb ≤ 7 g/dl) and liberal transfusion (Hb ≤ 9.5 g/dl) in stable hemodynamic critically-ill children showed no significant differences. Other studies by Cholette et al. 7 compared these two transfusion strategies in children with cyanotic heart disease also reported no significant differences in circulation parameter. In pediatric sepsis population, early RBC transfusion may benefit to reduce multiple organ dysfunction days in more severe stage of septic shock. 8 Furthermore, it remains questionable which the predictor for outcome is, the initial, increment, or final Hb level. The primary objective of this study was to compare restrictive and liberal transfusion strategy; whereas the secondary objectives were to compare Hb increment and final Hb level to clinical outcome of critically ill children.

| METHODS
This is an observational retrospective study. Subjects was children aged 1 month-18 years old who had been hospitalized at high care unit (HCU) or PICU, and received PRC transfusion. Patient who had any hematologic or congenital disorder, severe malnutrition, chronic infection-related anemia, and transfusion in Hb level ≥ 9.5 g/dl were excluded from this study. Subjects categorized into two groups based on initial Hb level, that is restrictive (Hb ≤ 7 g/dl) and liberal (Hb ≤ 9.5 g/dl) transfusion strategy. Then, in each group, subjects categorized based on: (1) Hb increment and (2) final Hb level. First, based on Hb increment, subjects divided into two groups, that is high (increased ≥2.5 g/dl) and low increment (increase <2.5 g/dl). Second, by the final Hb level, subject divided into three groups, low   (Table 1).  Table 3).

| Mortality rate
In condition like hipoxemia or shock, the increase of oxygen demand cannot be accompanied by DO 2 raise. To maintain the cellular oxygen uptake, the compensatory mechanism is to increase the extraction rate of oxygen (ErO 2 ). This compensation had a limit and in the critical point the extraction rate reaches its peak and cannot compensate less DO 2 to maintain oxygenation. This condition happens in septic shock, when increase of metabolic demands and disruption of oxygen extraction causes inadequate tissue oxygenation. 10 The main aim of transfusion is to improve tissue oxygenation, which in turn, maintain organ perfusion and function. In the critical patient, improvement of tissue oxygen delivery can be beneficial to accelerate ventilator weaning. 9 In this study, In ventilator usage patient, the transfusion is succees if there is an increase of CO gained from preload addition. The weaning process will limitate the tissue oxygen delivery and the compensation is to increase the respiratory muscle work. This process gives an T A B L E 1 Baseline characteristics.

Characteristics
Restrictive (n = 65) Liberal (n = 120) failure. Some factor contribute to the fall of Hb level was the bleeding process and cristalloid infusion throughout surgery. 13,14 According to WHO, the recommendation of PRC transfusion volume is on 20 ml/kgBW, which can be predicted to increase Hb level for 2.5-3.3 g/dl. This amount of transfusion could minimize complication risk, which of caused by excessive transfusion. 15 There are some limitations in this study, considering the research design of analytic observational depends on patient's medical records. There is no data for Hb level on prehospital situation, so there is a chance of bias for patient's anemia condition. We also did not include the volume of transfusion in the data, thus the differences of clinical outcome which may be elicited by the amount of transfusion not being analyzed in this study. Future study with more samples or clinical trial design is needed to define the better transfusion for critical illness children in various disease and baseline condition.