Description of the condition

Neonatal hyperbilirubinaemia is one of the most common conditions confronting neonatologists. Approximately 60% of term and 80% of preterm infants develop hyperbilirubinaemia in the first week of life (Rennie 2010). Extremely high levels of serum bilirubin can lead to bilirubin encephalopathy, a devastating brain injury that can cause permanent neurodevelopmental handicaps. Neonatal phototherapy has been widely adopted as the initial therapy of choice for hyperbilirubinaemia. However, studies have shown a possible relationship between phototherapy and side effects such as interference with maternal–infant interaction, imbalance of the thermal environment and water loss, hypocalcaemia, disorder of circadian rhythms, and bronze baby syndrome (Xiong 2011). In hyperbilirubinaemia due to haemolytic disease, phototherapy has no effect on the levels of maternal antibody in the infant and on correcting the associated anaemia.

The terms 'kernicterus' and 'bilirubin encephalopathy' are used to describe the neurological consequences of extreme hyperbilirubinaemia. Although kernicterus is a pathological term, it is currently used interchangeably with bilirubin encephalopathy (Shapiro 2011). Infants with severe hyperbilirubinaemia may develop acute bilirubin encephalopathy, a clinical syndrome of lethargy, hypotonia and poor suck that may progress to hypertonia (with opisthotonos and retrocollis) with a high‐pitched cry and fever, and eventually to seizures and coma. Chronic bilirubin encephalopathy is defined as the clinical sequelae of acute encephalopathy with athetoid cerebral palsy (involuntary movements of the neck, face, arms and upper body) with or without seizures, developmental delay, hearing deficit, oculomotor disturbances, dental dysplasia and mental deficiency. The neurological findings correspond to the neuropathological lesions in the basal ganglia, auditory brainstem nuclei and perhaps the auditory nerve and brainstem oculomotor nuclei (Shapiro 2005).

Clinical signs of acute bilirubin encephalopathy may be absent in neonates who later develop bilirubin encephalopathy (Bhutani 2006). Auditory evaluation may improve detection of bilirubin‐induced neurotoxicity in neonates. Although sensorineural hearing loss has been widely reported in chronic bilirubin encephalopathy, more recently auditory dys‐synchrony (auditory neuropathy spectrum disorder), an auditory disorder characterized by a normal otoacoustic emission test but abnormal or absent auditory brainstem evoked response (ABR), has been described in early childhood in association with neonatal jaundice (Saluja 2010). This newly described term, 'auditory neuropathy', also called 'auditory dys‐synchrony', is functionally defined as absent or abnormal ABR with normal tests of inner ear function. In most cases of auditory neuropathy, ABRs performed at 80 dB are usually limited to early and fast waves (cochlear microphonics) that exhibit fixed‐latency function and complete phase reversal between rarefaction and condensation stimuli. In some instances of auditory neuropathy spectrum disorder, the ABR may show a wave V (a rapid decrement in amplitude following a peak that represents activity emanating from regions representing the rostral pons to midbrain) but with decreased amplitude and increased latency.

In the late 1940s, exchange transfusion was first introduced for newborns with rhesus haemolytic disease. The use of exchange transfusion was subsequently expanded to other causes of hyperbilirubinaemia as well as other neonatal diseases including sepsis, disseminated intravascular coagulation, severe fluid or electrolyte imbalance, polycythaemia, and severe anaemia. In early clinical studies on neonatal hyperbilirubinaemia, exchange transfusion was performed at an unconjugated bilirubin level > 342 µmoles/L. In 1994, the American Academy of Pediatrics guidelines increased the bilirubin threshold for initiating exchange transfusion in term infants without haemolysis (Steiner 2007). The development and widespread use of Rh‐immunoglobulin, improvements in diagnostic prenatal ultrasound, and intensive phototherapy have resulted in a worldwide decrease in the need for exchange transfusion during the last two to three decades. Currently, the estimated use of exchange transfusion in the United States is approximately 3/100,000 live births (Murki 2011). Although the use of exchange transfusion in developed countries is now rare, exchange transfusion remains a frequent procedure in developing countries, especially in Asian countries with a high incidence of neonatal hyperbilirubinaemia (Gharehbaghi 2010). This is because of suboptimal or ineffective phototherapy devices, delayed recognition of excessive bilirubin levels, and higher prevalence of glucose‐6‐phosphate dehydrogenase (G‐6‐PD) deficiency (Murki 2011). Arbitrary relaxation of exchange transfusion criteria has resulted in a recent resurgence of kernicterus (Ahlfors 2004). To date, severe neonatal hyperbilirubinaemia, especially haemolytic diseases, remains the most frequent indication for exchange transfusion (Murki 2011). Exchange transfusion is the only choice for prevention of kernicterus when phototherapy has failed (Mills 2001).

The effect of blood exchange is generally related to the recipient’s blood volume (Gharehbaghi 2010). Estimating the blood volume of the newborn as 80 to 90 mL/kg, exchange involving replacement of 80 to 90 mL/kg has been termed single volume exchange transfusion and that involving 160 to 180 mL/kg as double volume exchange transfusion. A single volume blood exchange transfusion is likely to exchange 63% of the infant’s blood whereas double volume blood exchange transfusion exchanges 86%. However, because of tissue to intravascular dynamics, more bilirubin is removed from the infant than the blood volume exchanged during exchange transfusion. A double volume blood exchange transfusion results in a maximal removal of bilirubin. A further increase in exchange volume did remove more bilirubin but to a much lesser extent (Murki 2011).

Exchange transfusion can cause some serious side effects. One review of the research found that up to one in 10 babies had some serious reactions (for example, internal bleeding). Between three in 1000 and four in 1000 babies who had an exchange transfusion died (Ip 2004). The causes of death ascribed to exchange transfusion include cardiovascular collapse during the transfusion and the subsequent complications of necrotizing enterocolitis, bacterial sepsis, and pulmonary haemorrhage. The most common events due to exchange transfusion are thrombocytopenia, hypocalcaemia, and metabolic acidosis (Patra 2004).

Description of the intervention

Human albumin, a protein with a molecular weight of 66,500 Daltons, is synthesized exclusively in the liver and is degraded in muscles, the liver, and the kidneys (Garcovich 2009). It is the most abundant protein in plasma (Caironi 2009; Rhee 2011). The total body albumin content is 4 to 5 gm/kg, of which one third is in the intravascular space and two thirds in the extravascular compartment (Farrugia 2010).

Albumin has been used clinically for over 50 years (Myburgh 2009). It was first fractionated during World War II for use as a volume expander (Rhee 2011). In adults, albumin infusions have been used in the management of patients with decompensated cirrhosis to reduce the formation of ascites (Rena 2010). Other novel uses have been reported, such as human albumin in plasma exchange as an approach for Abeta mobilization in Alzheimer's disease (Boada 2009). In infants, intravenous albumin infusion is used to treat hypoalbuminaemia, which occurs in infants born preterm, with respiratory distress syndrome, chronic lung disease, necrotizing enterocolitis, intracranial haemorrhage, hydrops fetalis, and oedema (Jardine 2004).

In 1959, Odell et al (Odell 1959) found that albumin administration could cause a shift of bilirubin from the extravascular to the intravascular compartment. Subsequent studies have shown that pre‐exchange albumin infusion improved the efficiency of exchange transfusion, as measured by the bilirubin removed, by up to about 40% (Odell 1959; Comley 1968; Wood 1970). Pre‐exchange albumin infusion has been a treatment of choice for neonates with severe hyperbilirubinaemia.

Recent controlled trials have demonstrated that albumin could benefit hyperbilirubinaemic neonates. In term non‐haemolytic hyperbilirubinaemic neonates during intensive phototherapy, there was a significant reduction in the serum unbound bilirubin values after albumin 1 gm/kg bodyweight (Hosono 2001) which was responsible for decreasing the rate of auditory brainstem response abnormalities at six months (Hosono 2002).

How the intervention might work

Albumin has several biochemical properties including buffering the pH; regulation of the colloid osmotic pressure of plasma; transportation of hormones, fatty acids, drugs, and metabolites in plasma; regulation of microvascular permeability; antioxidant activity; antithrombotic activity; and anti‐inflammatory activity (Prajapati 2011; Rhee 2011). Three functions of albumin are especially important; albumin is an important binding protein, the chief protein maintaining oncotic pressure, and a scavenger of oxygen radicals and other oxidizing agents (Roche 2008; Soeters 2009).

Albumin serves as a transport protein for bilirubin. Approximately 8.3 mg bilirubin is bound to each gram of albumin. Bilirubin entry into the brain is facilitated by drug displacement of bilirubin from its albumin binding site, reduced albumin binding capacity, and other factors (Hansen 2001). Thus free bilirubin, which is not bound to albumin, has been regarded as a parameter for bilirubin neurotoxicity (Hulzebos 2008). In this regard, it is evident that albumin concentration is associated with bilirubin neurotoxicity (Shahian 2010). Theoretically, in the case of hypoalbuminaemia there may be an increase in the ‘free’ unbound bilirubin and increased risk of bilirubin encephalopathy. This is especially important in preterm infants because of common hypoalbuminaemia (Bunt 2007). From a therapeutic viewpoint, pre‐exchange albumin infusion may be advantageous because an increased reserve of albumin may be protective against bilirubin toxicity by providing more binding sites, thereby reducing the levels of unbound bilirubin (Murki 2011; van Imhoff 2011). A rapid reduction in serum free bilirubin may be theoretically effective for the prevention of bilirubin encephalopathy.

There is an equilibration of bilirubin between plasma and the extravascular space. Following albumin administration, vascular bilirubin‐albumin binding would cause a shift of bilirubin from the extravascular to the intravascular compartment (Odell 1959). Pre‐exchange albumin infusion could provide more binding sites for bilirubin, leading to a shift of bilirubin from the tissues to the circulation. Therefore, more bilirubin would be removed through exchange transfusion, leading to an improved efficiency of exchange transfusion and decreased post‐exchange bilirubin levels (Odell 1959; Comley 1968; Wood 1970; Shahian 2010).

In addition, inflammatory processes, calcium overload, and oxidative damage contribute to the neural toxicity of bilirubin (Ostrow 2004; Watchko 2006; Fernandes 2009). Thus, albumin may protect the immature nervous system from bilirubin toxicity owing to its ability to scavenge metal ions, its anti‐inflammatory activity, and antioxidant properties (Prajapati 2011).

Why it is important to do this review

As an adjuvant drug, albumin has been widely used in neonates with jaundice, especially in infants needing exchange transfusion. However, the effects of pre‐exchange albumin infusion for unconjugated hyperbilirubinaemia in neonates requiring exchange transfusion are still controversial (Boldt 2010; Murki 2011). Whether albumin therapy could benefit the neonate with severe hyperbilirubinaemia and improve long‐term developmental outcomes remains unknown. Different methods of albumin separation have increased the variability of products between different manufactures, which affects albumin’s physiologic properties and therapeutic effects (Farrugia 2010). Albumin is a scarce and expensive resource with substantial cost. As a blood product, questions have been raised about its safety including the potential for allergic reactions and transmission of infection (Myburgh 2009). Fluid overload is another potential side effect of albumin administration (Jardine 2004). Therefore, it is necessary to perform this systematic review, searching for all the randomised controlled trials of pre‐exchange albumin infusion therapy for neonates with hyperbilirubinaemia, to summarize currently available evidence and to determine the benefits and risks of pre‐exchange albumin infusion.