Secondary Logo

Journal Logo


Anaesthetic strategies to reduce perioperative blood loss in paediatric surgery

Weber, T. P.; Hartlage, M. A. Groeß; Van Aken, H.; Booke, M.

Author Information
European Journal of Anaesthesiology (EJA): March 2003 - Volume 20 - Issue 3 - p 175-181
  • Free


Preoperative autologous blood donation

In adults, a number of measures to reduce perioperative blood loss have been established [1]. Preoperative autologous blood donation is the most effective method to reduce the exposure to homologous blood components in adult surgery [2]. However, it is far more invasive and complicated in infants or children:

• Infants may require light anaesthesia for blood donation. An inhalational anaesthetic technique is preferable; intravenous (i.v.) hypnotics, narcotics and muscle relaxants should be avoided to prevent any drug-related side-effect of the autologous blood unit after retransfusion.

• Regular blood collection bags cannot be used because they contain too much anticoagulant for the small amount of blood withdrawn.

• Mayer and colleagues found preoperative autologous blood donation impossible to perform in approximately 10% of their patients due to venous access problems [3].

Despite these difficulties most parents are enthusiastic since, in their opinion, the potential benefit from prevented homologous blood transfusion out-weighs the risk of additional anaesthesia and stress for their children. In adult surgery much preoperatively donated blood is not used intraoperatively and is thus discarded. Given the complexity of preoperative autologous blood donation in children and infants, its indication should be limited to patients with a very high likelihood of subsequent transfusion requirement.


Perioperative haemodilution is much easier to perform in children than is preoperative blood donation since no additional anaesthesia is needed to withdraw the blood. However, the efficacy of haemodilution, surprisingly, has not been established or even questioned in adults. A recent meta-analysis showed that haemodilution in randomized controlled clinical trials with clearly established transfusion triggers did not reduce the exposure to homologous blood [4]. However, several studies included in that meta-analysis did not consider the basic principles of haemodilution. These include a high initial haematocrit (>35%), and the tolerance of a low haematocrit (<25%) intraoperatively. If these criteria are not fulfilled, the amount of blood withdrawn is insufficient to make a difference in exposure to homologous blood.

In paediatric surgery, these 'entry criteria' often cannot be attained. In children and infants, an increase in oxygen delivery during haemodilution cannot be achieved by an increase in stroke volume (Fig. 1). Furthermore, during the first 4-6 months of age, haemodilution is contraindicated due to the high percentage of fetal haemoglobin. Fetal haemoglobin is characterized by a leftward shift of the oxygen dissociation curve, which limits oxygen availability in peripheral tissues. Low haematocrits at this stage may put the infant at risk of organ ischaemia and hypoxia. Infants above the age of 4-6 months are characterized by a physiologically low haematocrit, making haemodilution also inapplicable. Thus, haemodilution is only indicated in patients being ≥1 yr of age (Figs 2 and 3). In those patients, it may well be a useful method to reduce homologous transfusion requirements [5,6] although controlled randomized clinical trials are still lacking. Recently, a Japanese group investigated the effectiveness of intraoperative autologous blood donation from arterial monitoring lines during open-heart surgery without the use of homologous blood in infants and young children. They performed haemodilution in two stages in a group of 81 patients, weighing 5.5-14.9 kg. First, following induction of anaesthesia, approximately 10 mL kg−1 blood was withdrawn. These haemodiluted infants were then connected to a cardiopulmonary bypass circuit, which had been primed with clear fluids alone, without the addition of any red blood cells. No adverse events were noted [7].

Figure 1
Figure 1:
In children and infants, an increase in oxygen delivery during haemodilution cannot be achieved by an increase in stroke volume.
Figure 2
Figure 2:
Fetal haemoglobin with its higher oxygen affinity may limit peripheral oxygen consumption during haemodilution. SaO2: blood oxygen saturation; PO2: oxygen partial pressure.
Figure 3
Figure 3:
The physiological low haematocrit does not allow for sufficient withdrawal of autologous blood. Hct: haematocrit.

Intraoperative autotransfusion

Intraoperative autotransfusion has recently proved to be highly effective not only in terms of reducing blood loss, but also in terms of reducing exposure to homologous blood [8]. This also holds true when shed blood is retransfused without being previously washed. However, these positive results cannot simply be applied directly to paediatric surgery for the following reasons:

• The absolute amount of shed blood in infants is rather small although relatively high for the patients themselves. These small volumes of shed blood are hard to collect because most of the blood is not aspirated into a reservoir but rather collected on sponges and swabs. Although these sponges and swabs can be washed in saline, up to 50% of the red blood cells lost intraoperatively cannot be recovered. This is far less than in adults.

• In order to wash shed blood effectively with an autotransfusion device, a certain amount of blood has to be collected before the washing procedure can be initiated. Until recently, this minimum amount of shed blood was approximately 300 mL [9].

• This minimum blood loss of about 300 mL must be haemodynamically tolerated by the infant. Small infants or neonates certainly will not tolerate a blood loss of 300 mL without receiving homologous blood. At this point, the aim to prevent any exposure to homologous blood can no longer be fulfilled. However, autotransfusion may reduce the number of homologous blood units transfused perioperatively. Given a normal preoperative haematocrit, and under the assumption that at least 50% of the shed blood can be collected, children above approximately 20 kg body weight should tolerate a blood loss of 300 mL. In children <20 kg body weight, conventional intraoperative autotransfusion is not applicable [10].

• The indication for intraoperative autotransfusion could be broadened if the attending anaesthesiologist tolerates a lower haematocrits in infants. Currently, most experts recommend keeping the haematocrit >30% in infants. This is far more than that recommended for otherwise healthy adults. The reason relates to the high proportion of fetal haemoglobin within the first 6 months of life. Fetal haemoglobin is characterized by a high affinity for oxygen that, in turn, limits oxygen supply to the peripheral tissue [11]. Therefore, lower haematocrits during the first 6 months may be harmful since oxygen delivery is reduced owing to the low haematocrit while oxygen supply to the peripheral tissue is reduced owing to the fetal haemoglobin. Older children are not affected by these limitations and haematocrits far below 30% are tolerated by these patients. Messmer's group has studied 16 patients with body weights between 11 and 23 kg and aged between 12 and 89 months [12]. They defined a haematocrit of 17% as their transfusion trigger. Global tissue oxygenation was not compromised in any of these patients despite their low haematocrits. The acceptance of such low haematocrits, in turn, resulted in an increase of the maximal tolerable blood loss before the transfusion of homologous blood was indicated. Therefore, the above-mentioned minimum blood loss of 300 mL may be tolerated even in smaller children and, consequently, intraoperative autotransfusion may be indicated in children with a body weight <20 kg.

All conventional autotransfusion devices process blood in batches whereby the bowl initially is filled with shed blood from the reservoir. During this filling process, the bowl rotates at approximately 5000 rpm. The resulting centrifugal forces cause the red blood cells to move to the outer part of the bowl while particles of lighter density (plasma) are found in the centre. These lighter particles are then pumped into the waste bag. Through this process, the red blood cells are concentrated to a target haematocrit of 50-60%. The blood is then washed with saline and finally the washed blood is pumped into the retransfusion bag. This procedure is only effective as long as the bowl is completely filled with red blood cells at a haematocrit of 50-60%. For a conventional bowl with a capacity of 125 mL, approximately 300 mL blood with a haematocrit of 25% are needed to fill this bowl completely. This is another reason for the above-mentioned minimum blood loss.

Two technical advances that have been introduced recently both serve to broaden the indication for autotransfusion to smaller children:

• A bowl with an extremely small capacity (55 mL) has been introduced (Dideco Compact A & Advance®; Sorin Biomedica, Puchheim, Germany). It can be filled as soon as approximately 100 mL shed blood is available. However, the disadvantage of this technology is that if the bowl is not completely filled, a unit of autologous blood with an unpredictable haematocrit and an unpredictable quality is achieved.

• Continuous autotransfusion device (CATS®; Fresenius AG, Bad Homburg, Germany) is the first device not based on a (Latham) bowl. Shed blood is pumped into a separation chamber (capacity approximately 30 mL) while simultaneously saline 0.9% is added for the washing procedure. The final product is then pumped into a retransfusion bag. This device allows the effective processing of shed blood even if <100 mL are available. The delivered autologous blood product is characterized by a constant quality and haematocrit independent of the amount of shed blood. This, and the higher recovery rate of red blood cells, makes the CATS® superior when compared with the Dideco® [13].

Because both devices allow the processing of small blood volumes, these techniques are applicable to paediatric surgery. They can be used in paediatric surgery even when infants <10 kg body weight are undergoing surgery [14]. In these patients, the Dideco® and the CATS® may serve to avoid transfusion of allogeneic blood. However, it will take some time for this technology to overcome the established contraindication of <300 mL blood loss for the use of intraoperative autotransfusion.

In paediatric cardiac surgery, the volume remaining in the extracorporeal circulation after bypass cannot be retransfused completely because of the potential for volume overload. Friesen and colleagues effectively reduced exposure to homologous blood by ultrafiltration of the extracorporeal circuit volume before retransfusion [15]. Blood was concentrated to a haemoglobin >20 g dL−1, but free haemoglobin and heparin were also concentrated to undesirably high values. As an alternative, washing the residual volume with an autotransfusion device will also result in haemoglobin concentrations >20 g dL−1 with nearly complete elimination of free haemoglobin and heparin. This technique therefore seems superior to ultrafiltration.

The retransfusion of shed blood, which has not been washed before retransfusion, does not have these limitations. It can easily be performed independently of the amount of shed blood. However, this strategy cannot be recommended for paediatric surgery. Retransfusion of unwashed shed blood is known to cause an activation of the fibrinolytic system. This may produce fibrinolysis of wound clots resulting in an immediate increase in blood loss [16]. No data are available concerning the retransfusion of unwashed shed blood in infants.


During cardiopulmonary bypass an abnormal activation of the coagulatory and fibrinolytic system occurs. While coagulation is prevented by intravenous heparin, the impact of the activated fibrinolysis has long been overlooked. In the extracorporeal circulation, contact activation of the blood stimulates plasminogen activator, which in turn activates plasminogen, initiating fibrinolysis with subsequent bleeding [17].

In current practice, antifibrinolytic agents are frequently used prophylactically to reduce perioperative bleeding in patients undergoing open-heart surgery. A recent meta-analysis showed that aprotinin, as well as certain lysine analogues (tranexamic acid, epsilon-aminocaproic acid), significantly reduced blood loss, which translated to a significantly reduced exposure to homologous blood components [18]. Concurrently, the frequency of surgical re-exploration has also been shown to be significantly lower in patients receiving antifibrinolytic agents.

These results are not transferable to paediatric cardiac surgery. The balance between coagulation and fibrinolysis is far more delicate and susceptible to exogenous stimulation. Plasma concentrations of the vitamin K-dependent clotting factors (II, VII-X), proteins S and C, and the components of the contact system (Hageman factor, prekallikrein, kininogen, Factor XI) are all lower in neonates and small infants, probably because of decreased hepatic synthesis [19]. Cyanotic patients further demonstrate impaired haemostasis related to polycythaemia, low platelet count, abnormal platelet function, decreased concentrations of Factors V, VII and VIII, and increased fibrinolysis, all of which are changes that directly correlate with the degree of cyanosis [20].

The conduct of cardiopulmonary bypass, including extensive cooling and, in many cases, cardiac arrest and the complex intracardiac surgical procedures performed, also disadvantage the paediatric population compared with the adult population. Furthermore, the adverse size relationship between the patient and the heart-lung machine dictates that more profound haemodilution occurs with undesirable dilution of clotting factors which may already be reduced.

Since haemostasis in children is not as robust as in adults, and while the stimulus of coagulation and fibrinolysis is even worse than in adults, blocking coagulation and fibrinolysis seems to be even more important in paediatric cardiac surgery. Nonetheless, the results of aprotinin as an antifibrinolytic agent in paediatric cardiac surgery are conflicting. Davies and colleagues found aprotinin not to be effective in routine paediatric cardiac surgery [21]. Coniff found a trend (not achieving statistical significance) toward benefit with aprotinin use in a paediatric population >1 yr undergoing a repeat cardiac operation [22]. However, children <1 yr of age as well as those undergoing primary operation showed no benefit. Carrel and colleagues looked at 168 children with a body weight <15 kg [23]. They found a high-dose aprotinin regimen (50 000 KIU kg−1 before initiation of the extracorporeal circuit, 50 000 KIU kg−1 in the pump prime and 20 000 KIU kg−1 continuous infusion) to be effective only in complex cardiac operations. One reason for the wide variability in aprotinin efficacy is the significantly differing doses. For paediatric cardiac surgery, the dose should be adjusted to the body surface area rather than to the weight, which in neonates results in a 2.5 times higher dose. In addition, aprotinin should be added to the pump-prime volume to compensate for the often extreme haemodilution. If the dose of aprotinin given is insufficient, the resulting plasma concentrations are <200 KIU mL−1 (0.03 mg mL−1), which will not prevent the above-mentioned contact activation during bypass.

The same concepts are similarly applicable to the use of tranexamic acid and epsilon-aminocaproic acid. Both compounds have proved to reduce blood loss effectively if the given doses are high enough to compensate for the haemodilution caused by the pump-prime volume [24,25].


During the early phase of coagulation, platelets bind via the glycoprotein receptor Ib to the von Willebrand factor, which is exposed by damaged endothelium (Fig. 4). Desmopressin given i.v. results in a release of von Willebrand's factor from endothelial storage pools and also enhances the expression of glycoprotein receptors on the platelet surface. Thus, the binding of platelets to damaged endothelium can be enhanced. The maximum effect can be obtained at a dose of 0.3 μg kg−1 body weight [26]. To prevent platelet activation during extracorporeal bypass, desmopressin should not be given prophylactically before bypass (like antifibrinolytics) but should be given after termination of the extracorporeal circulation [27].

Figure 4
Figure 4:
Platelet interaction with the damaged endothelial surface. VW factor: von Willebrand factor; FAC: FAC fragment.

In 1986, Salzman and colleagues showed that in adults undergoing high-risk cardiac surgery the administration of desmopressin after cessation of the extracorporeal circulation resulted in a significant reduction in blood loss [28]. Since then, multiple studies have been undertaken in adult cardiac surgery. However, a recent meta-analysis showed that desmopressin has no effect on perioperative blood loss during routine cardiac surgery although it significantly reduced blood loss during high-risk surgery, such as repeat cardiac surgery or surgery involving patients with recent aspirin intake [29].

Data on desmopressin during paediatric cardiac surgery are very limited but conclusive. Seear and colleagues looked at the effects of desmopressin (0.3 μg kg−1) given after termination of the extracorporeal circulation in paediatric patients undergoing open heart surgery [30]. Desmopressin had no effect on perioperative blood loss. However, the studied patient population could not be classified as high risk. Given these data and the publications showing desmopressin to be ineffective in routine cardiac surgery in adults, Reynolds and colleagues studied high-risk cardiac surgery in children <2 yr of age [31]. In both study groups, deep hypothermic cardiac arrest was used in 97% of patients. They could not find any difference in perioperative blood loss between groups and found further that von Willebrand's factor increased to the same extent no matter whether or not desmopressin was given. This is in agreement with Gill and Montgomery who showed that compared with older children, younger children are not as capable of releasing von Willebrand's factor from endothelial storage sides [32]. Obviously, the operative stimulus causes a maximum release of von Willebrand's factor that cannot be further enhanced by desmopressin. Consequently, the use of desmopressin to reduce perioperative blood loss cannot be recommended in paediatric surgery.

Minimal exposure transfusion

When a neonate or infant requires transfusion it is a practical point to increase the haematocrit to the upper limit of normal. This will decrease the likelihood of subsequent transfusion possibly using blood from another donor. Whenever possible, regular blood units should be divided into several (e.g. six) sub-units. After transfusion of the first sub-unit, all remaining sub-units should be reserved for this particular patient. This is one of the reasons why blood for neonates and infants should not be older than 1 week. The other reason for this recommendation is that older red blood cells are depleted of 2,3-diphosphoglycerate (2,3-DPG) with a subsequent leftward shift of the oxygen dissociation curve, which, especially in neonates, may aggravate peripheral tissue hypoxia (as discussed above).

Another strategy to reduce transfusion-related immunological and/or infectious complications is to use only a minimum number of donors who contribute all homologous blood and blood components that a specific patient is expected to need. This approach has also been coined 'minimal exposure transfusion' [33]. This technique is most effective in neonates and infants. Although the blood loss may be considerable for these patients, the absolute amount of blood lost is still low. Thus, one committed donor can contribute virtually all blood components, including plasma and platelets, needed for these small patients.


1. Booke M, Hinder F, Van Aken H, et al. Methods to reduce postoperative blood loss - Verfahrenstechniken zur Reduktion des postoperativen Blutverlusts. Anesth Intensivmed 1999; 40: 535-540.
2. Forgie MA, Wells PS, Laupacis A, Fergusson D. Preoperative autologous donation decreases allogeneic transfusion but increases exposure to all red blood cell transfusion: results of a meta-analysis. International Study of Perioperative Transfusion (ISPOT) Investigators. Arch Intern Med 1998; 158: 610-616.
3. Mayer MN, de Montalembert M, Audat F, et al. Autologous blood donation for elective surgery in children weighing 8-25 kg. Vox Sang 1996; 70: 224-228.
4. Bryson GL, Laupacis A, Wells GA. Does acute normovolemic hemodilution reduce peroperative allogeneic transfusion? A meta-analysis. The international study of perioperative transfusion. Anesth Analg 1998; 86: 9-15.
5. Velardi F, Di Chirico A, Di Rocco C, et al. 'No allogeneic blood transfusion' protocol for the surgical correction of craniosynthesis. I. Rationale. Childs Nerv Syst 1998; 14: 722-731.
6. Velardi F, Di Chirico A, Di Rocco C, et al. 'No allogeneic blood transfusion' protocol for the surgical correction of craniosynthesis. II. Clinical application. Childs Nerv Syst 1998; 14: 732-739.
7. Kurisu K, Yonenaga K, Miyamoto K, Furusho N, Nishimura K. Open heart surgery without the use of homologous blood in infants and young children: effectiveness of the intraoperative autologous blood donation from arterial monitoring line. Kyobu Geka 1999; 52: 438-441.
8. Huet C, Salmi LR, Fergusson D, Koopman-van Gemert AW, Rubens F, Laupacis A. A meta-analysis of the effectiveness of cell salvage to minimize perioperative allogeneic blood transfusion in cardiac and orthopedic surgery. International study of perioperative transfusion. Anesth Analg 1999; 89: 861-869.
9. Michaelis G, Melzer C, Biscoping J, Hempelmann G. Mechanical autologous transfusion in orthopedic surgery in children. Is the use of mechanical autologous transfusion possible even in pediatric orthopedic surgical procedures? Anaesthesist 1995; 44: 501-507.
10. Mazzarello G, Lampugnani E, Carbone M, Rivabella L, Ivani G. Blood saving in children. Anaesthesia 1998; 53: 30-32.
11. Schwarz U. Intraoperative fluid therapy in infants and young children. Anaesthesist 1999; 48: 41-50.
12. Aly Hassan A, Lochbuehler H, Frey L, Messmer K. Global tissue oxygenation during normovolemic haemodilution in young children. Paediatr Anaesth 1997; 7: 197-204.
13. Booke M, Hagemann O, Van Aken H, et al. Intraoperative autotransfusion in small children: an in vitro investigation to study its feasibility. Anesth Analg 1999; 88: 763-765.
14. Booke M, Van Aken H. What is the minimum weight required to effectively perform intraoperative autotransfusion in small children - an in vitro investigation. Anaesth Intensive Care 2001; 29: 74-77.
15. Friesen RH, Tornabene MA, Coleman SP. Blood conservation during cardiac surgery: ultrafiltration of the extracorporeal circuit volume after cardiopulmonary bypass. Anesth Analg 1993; 77: 702-707.
16. Riou B, Arock M, Guerrero M, et al. Haematological effects of postoperative autotransfusion in spinal surgery. Acta Anaesthesiol Scand 1994; 38: 336-341.
17. Furie B, Furie BC. Molecular and cellular biology of blood coagulation. N Engl J Med 1992; 326: 800-806.
18. Levi M, Cromheecke ME, de Jonge E, et al. Pharmacological strategies to decrease excessive blood loss in cardiac surgery: a meta-analysis of clinically relevant endpoints. Lancet 1999; 354: 1940-1947.
19. Andrew M, Paes B, Johnston M. Development of the hemostatic system in the neonate and young infant. Am J Pediatr Hematol Oncol 1990; 12: 95-104.
20. Ekert H, Gilchrist GS, Stanton R, Hammond D. Hemostasis in cyanotic congenital heart disease. J Pediatr 1970; 76: 221-230.
21. Davies MJ, Allen A, Kort H, et al. Prospective, randomized, double-blind study of high-dose aprotinin in pediatric cardiac operations. Ann Thorac Surg 1997; 63: 497-503.
22. Coniff RF. The Bayer 022 compassionate-use pediatric study. Ann Thorac Surg 1998; 65: S31-S33.
23. Carrel TP, Schwanda M, Vogt PR, Turina MI. Aprotinin in pediatric cardiac operations: a benefit in complex malformations and with high-dose regimen only. Ann Thorac Surg 1998; 66: 153-158.
24. Willimas GD, Bratton SL, Riley EC, Ramamoorthy C. Efficacy of epsilon-aminocaproic acid in children undergoing cardiac surgery. J Cardiothorac Vasc Anesth 1999; 13: 304-308.
25. Reid RW, Zimmerman AA, Laussen PC, Mayer JE, Gorlin JB, Burrows FA. The efficacy of tranexamic acid versus placebo in decreasing blood loss in pediatric patients undergoing repeat cardiac surgery. Anesth Analg 1997; 84: 990-996.
26. Lethagen S. Desmopressin - a haemostatic drug: state-of-the-art review. Eur J Anaesthesiol 1997; 14 (Suppl 14): 1-9.
27. Shiffrin JS, Glass DD. Pro: desmopressin is of value in the treatment of post-cardiopulmonary bypass bleeding. J Cardiothorac Vasc Anesth 1991; 5: 285-289.
28. Salzman EW, Weinstein MJ, Weintraub RM, et al. Treatment with desmopressin acetate to reduce blood loss after cardiac surgery. A double-blind randomized trial. N Engl J Med 1986; 314: 1402-1406.
29. Laupacis A, Fergusson D. Drugs to minimize perioperative blood loss in cardiac surgery: meta-analysis using perioperative blood transfusion as the outcome. The international study of perioperative transfusion (ISPOT) investigators. Anesth Analg 1997; 85: 1258-1267.
30. Seear MD, Wadsworth LD, Rogers PC, Sheps S, Ashmore PG. The effect of desmopressin acetate (DDAVP) on postoperative blood loss after cardiac operations in children. J Thorac Cardiovasc Surg 1989; 98: 217-219.
31. Reynolds LM, Nicolson SC, Jobes DR, et al. Desmopressin does not decrease bleeding after cardiac operation in young children. J Thorac Cardiovasc Surg 1993; 106: 954-958.
32. Gill JC, Montgomery RR. Failure of response to DDAVP in children less than one year of age. Pediatr Res 1986; 20: 279.
33. Brecher ME, Taswell HF, Clare DE, Swenke PK, Pineda AA, Moore SB. Minimal-exposure transfusion and the committed donor. Transfusion 1990; 30: 599-604.

BLOOD TRANSFUSION, autologous, blood transfusion; HAEMORRHAGE, blood loss, surgical; INFANT, NEWBORN

© 2003 European Society of Anaesthesiology