Albumin: saint or sinner?
- SIMON NADEL, Consultant in Paediatric Intensive Care
- CLAUDINE DE MUNTER, Consultant in Paediatric Intensive Care
- JOSEPH BRITTO, Consultant in Paediatric Intensive Care
- MICHAEL LEVIN, Professor of Paediatrics
- PARVIZ HABIB, Senior Lecturer in Paediatric Intensive Care
The choice of colloid or crystalloid solutions in adult and paediatric resuscitation has been the subject of ongoing debate for many years. A systematic review recently published in the BMJhas again brought this debate to the fore, by indicating that patients treated with human albumin solution (HAS) may have an increased mortality.1 The review examined the use of HAS for its licensed indications: hypoalbuminaemia; hypovolaemia from trauma or surgery; and burns. On closer analysis, it is evident that the main aim of many of the studies included was to compare haemodynamic variables and the effects of HAS on nutritive status, rather than to assess the effects of HAS on mortality. While two of the included studies evaluated HAS use in preterm infants, only one assessed HAS in sepsis and none evaluated treatment of critically ill children with HAS. The authors of the systematic review concluded that the use of HAS was associated with a 6% excess mortality, and that HAS use in critically ill patients should be reviewed. An expert working party is considering the implications of this data.
Where does this leave the practising clinician in the emergency management of the seriously ill or injured child?
The most common cause of childhood mortality worldwide is hypovolaemia, either from gastroenteritis and dehydration, sepsis or trauma. For patients with shock from any cause, treatment of the underlying cause is mandatory. Rapid intravascular volume expansion guided by clinical examination and urine output is frequently adequate to restore organ perfusion and blood pressure. In paediatric septic shock, Carcilloet al showed that fluid resuscitation, of at least 40 ml/kg in the first hour after presentation, was associated with improved survival without increased risk of pulmonary oedema or the acute respiratory distress syndrome.2
What fluid is available
Decisions regarding which fluid should be used for resuscitation of hypovolaemic children should consider the type of fluid lost, maintenance of plasma oncotic pressure, risk of infection, and cost. Less than 25% of administered crystalloid solution remains within the intravascular space in normal conditions. The remainder rapidly and freely fills the interstitial and intracellular fluid compartments.3-5 Therefore, four times the volume of crystalloid solution would be required to have an equivalent volume enhancing effect of a substance that remained wholly intravascularly.
Albumin is a plasma protein that provides approximately 80% of intravascular colloid oncotic pressure in normal subjects. Its molecular weight (MW) is 69 000 and it is negatively charged at physiological pH. Albumin is relatively impermeable to the endothelial barrier under normal conditions, where its intravascular half life is 24 hours, with haemodynamic improvement persisting up to 36 hours following administration.6 Albumin has the additional advantage that it is a physiological protein with many other functions including protein, hormone, and drug binding.
Maintaining plasma colloid oncotic pressure appears to be important in control of normal organ function. Studies in adults have found strong correlations between decreased plasma colloid oncotic pressure (caused by leakage of plasma colloids) and the subsequent development and severity of pulmonary oedema.7 8
Synthetic colloids are cheaper than HAS and without the theoretical risk of infectious complications. They are either gelatin or starch derivatives. The gelatin derived compounds (such as Haemaccel (Hoechst Marion Roussel, Middlesex, UK) and Gelofusine (B Braun Medical Ltd, Bucks, UK)) contain various sized molecules (MW 5000–50 000, mean 35 000), which do not have much benefit over crystalloids in conditions of increased vascular permeability.9 The starch based compounds (such as Pentastarch and Hetastarch (Geistlich Sons Ltd, Cheshire, UK)) while larger than albumin (approximate MW 250 000), expand the plasma volume to a similar degree.4 10-12 All the artificial colloids cause adverse effects such as haemostatic abnormalities and anaphylaxis.12 This risk is increased by the use of larger volumes.13
Treatment of hypovolaemia
For hypovolaemia caused by dehydration, the estimated fluid and electrolyte deficits should be replaced with crystalloid. For trauma, effective circulating volume is decreased because of blood loss and reduction in interstitial fluid volume. Although several types of fluid have been suggested for trauma resuscitation, almost any fluid is suitable for initial resuscitation. However, there is no substitute for blood, particularly if the child does not respond to the initial fluid bolus.13 14
Treatment of sepsis
The endothelial barrier is disrupted in sepsis causing a capillary leak of plasma proteins and water into the extravascular compartment. Initially this protein leak appears to be size and charge dependent: smaller molecules leak more and negatively charged molecules are preferentially retained within the intravascular compartment.15 However, as endothelial dysfunction progresses molecules of all sizes leak. In theory the consequences of large amounts of extravascular protein or any other osmotically active material, are increased extravascular water leading to tissue oedema, persisting until the extravascular colloid is removed or degraded.
Our clinical experience with use of 4.5% HAS as first line resuscitation fluid in septic shock indicates that it is an effective treatment with no evidence to suggest that it is associated with excess mortality.
Meningococcal septicaemia causes a profound capillary leak syndrome together with myocardial dysfunction and multisystem failure.16 Pulmonary oedema occurs in up to 20% of children with meningococcal sepsis because of the capillary leak and is often present even before any fluid resuscitation.
We have cared for 410 children with meningococcal disease in the past six years. The median volume of fluid resuscitation they required to restore circulating volume was 80 ml/kg (range 20–300) in the first 12 hours of admission. The vast majority of this fluid was composed of 4.5% HAS. Despite the use of such large volumes of 4.5% HAS, we have had a lower than predicted case fatality ratio.17 In addition, no child has died on our unit from pulmonary oedema secondary to meningococcal disease since 1993. While this was not a properly conducted study, the fact that mortality in our series of children with meningococcal disease is lower than any of the published figures does not suggest that 4.5% HAS is dangerous in this condition.
Studies that have compared colloid (both natural and artificial) and crystalloid solutions for resuscitation have found little difference between them despite the theoretical arguments put forward above.18-20 No study has specifically evaluated the optimal resuscitation fluid for children or adults with sepsis, and none has examined the use of large volumes of artificial colloids. Therefore, there are no data to guide us on the rational choice of resuscitation fluid for children with sepsis. Until properly conducted, controlled trials are carried out that compare resuscitation fluids in children with sepsis, 4.5% HAS should remain the first choice for this condition despite its cost and theoretical risks of transmission of infective agents. In addition, controlled studies will be required for other indications where fluid resuscitation is felt necessary to compare the use of crystalloids, artificial colloids, and albumin.