 |
|
Editorial Indian Pediatrics 2002; 39:619-624 |
|
Blood Transfusions in the Premature Nursery |
|
Irrespective of the source of blood however, the clinical problem of deciding when to transfuse preterm infants remains. The scale of the problem in preterm nurseries is large: recent data from Canada suggests that in addition to frequent transfusions, there is a great deal of variation amongst tertiary care neonatologists as to what levels of hemoglobin merit transfusion(3,8). Minimizing Potential Need For Transfusions The clinical dilemma inevitably largely results from the necessary frequent blood sampling, resulting from intensive care(9). However both malnutrition and ineffective erythropoiesis may also play a role in the aetiology(10). Could requirements for transfusion be diminished by any attention to these three issues - sampling, nutritional supplementation, or erythropoietin? Attempts to limit blood sampling are fraught with a potential to overlook serious complications of care. Earlier, logistically awkward methods such as spinning down the discard and re-infusing the red cells were proposed(11). However a new technology proposes to circumvent this by the so-called ‘point-of-care testing’, whereby blood is withdrawn for checking and then is simply re-infused into the infant. A recent study by Widness and coworkers examined this in an RCT(12). Enrolling infants between 500-1000 g birth weight (BW), the aim was to reduce blood transfusions by some 35% using a system linked to an umbilical arterial catheter. The trial was successful. However it was curious that the goal was achieved only because of one center, while the other center had directly contradictory results. This un-blinded study is not therefore robust enough to change practice, without a more convincing multi-center demonstration, to justify the economic outlays in such expensive technology. Less expensive variants of this type of approach include attempts at minimizing loss of ‘discard volume’ by an adjustment of the withdrawal technique(13), which takes account of the apparent minimum need of a "draw-up" volume of 1.3 cc(14). An even simpler approach is to limit blood sampling and "phlebotomy overdraw" by attention to details such as design of blood collection containers(15). By and large an earlier enthusiasm for aggressive early intervention with erythropoietin(16) in order to bolster the bone marrow, has waned. A recent overview by Vamvakas and Strauss(17), found ‘only modest effects on overall needs’, defined as transfusion requirements. Most pertinent data in that review, drew upon a then preliminary report that itself found that erythropoietin with iron ‘does not reduce transfusion requirement’ (18). The overview(17) appropriately concluded further study was mandated. Subsequently, the full publication of the trial of Ohls et al, showed convincingly that there were no short term benefits to the most at risk infants (19). These were the 172 infants of less than 1000 g BW. Although erythropoietin-treated infants experienced the expected reticulocytosis, the overall effect on numbers of transfusions per infant was not clinically important and did not reach statistical significance (means 4.3 vs. 5.2). More important, there were no benefits attributable to erythropoietin use in major outcomes, such as survival or reduction in chronic lung disease, retinopathy or intraventricular haemorrhage. Ohls et al concluded, "the early use of erythropoietin and iron to reduce transfusion number and exposure is not warranted". They have since also shown that there are no long term benefits in terms of neurodevelopmental outcomes(20). Furthermore, any final residual pro-erythropoietin sentiments were probably chilled by the revelations of antibody to erythropoietin in various clinical settings, as reviewed recently by Zipursky(21). Thus far then, it may be concluded that minimal blood sampling, newer methods to minimize wastage of blood during the sampling process - and guidelines - maybe indicated. But we are still left with considerable uncertainty as to when it is best to transfuse. When to Transfuse? Can a more serious consideration of physiology assist us? These might weigh postulated benefits in terms of oxygen delivery to the tissues(22). The risks and benefits of transfusion include those of maintaining a high or low hemoglobin and some additional risks and benefits of the transfusion itself. A high hemoglobin level, maintained by frequent transfusion, enhances arterial oxygen content and oxygen transport to the tissues. But this is usually far in excess of need, and so oxygen delivery (equal to oxygen uptake or consumption) is not limited by hemoglobin content(23). However, in chronic ischemic or hypoxic hypoxia, where oxygen delivery may be limited by oxygen transport, a high hemoglobin may be required to maintain oxygen delivery to the tissues. Expected consequences of chronic anemic hypoxia might be thought to be poor growth or impaired neurodevelopmental outcome. On the other hand, if allowing the hemoglobin to fall to lower levels has no critical or limiting effects on oxygen delivery, growth and development will continue unimpaired without the potential adverse effects of blood transfusion, such as transfusion-borne infection or iron overload(5,24,25). Even further complicating these physiologic considerations is the decrease in oxygen affinity of hemoglobin with postnatal age, which increases the ability of the blood to deliver oxygen, and the effect of transfusion of adult hemoglobin, which enhances this effect. The net balance to the tissues might be oxygen available for extraction. Using this type of reasoning, the peripheral fractional oxygen extraction (FOE) may be a better indicator of the need for transfusion than the hemoglobin concentration because it is a measure of the adequacy of oxygen delivery to meet demand. To test this Wardle et al carried out a randomised controlled study in preterms less than 1500 g. They measured peripheral fractional oxygen extraction non-invasively using near infrared spectroscopy (NIRS). However they found that FOE measurements failed to identify many infants felt to be in need of transfusions by clinicians(26). As the authors identify themselves, this might be because clinicians are determining "need" for transfusion in a "vague" manner, or because the identification targets of an increased oxygen extradition are not correctly identifying infants in "need". Perhaps the solution would be to simply decree by fiat - intelligent fiat of course! - when to and when not to transfuse? This certainly removes variation, and reduces rates of transfusion, as shown for instance by Franz(27). Many other such guidelines exist, such as the Canadian concensus(28). While this achieves the benefit of reduced transfusions, it still remains uncertain as to whether there is any meaningful clinical trade off. In other words, are there risks that outweigh potential benefits by such guidelines? Bitter memories should lead neonatologists to recall Silverman’s dictum to enter into Randomized controlled trials(29). This leads then to a consideration of what randomized evidence is there in the newborn population that examines this question? What Clinical Data is Available? The clinical symptoms suggested as being caused by ‘anemia’ in the newborn traditionally ranges from apnea, higher oxygen requirements, longer ventilator durations, to slow weight gain(10). However when critically appraised by objective parameters, many of these are not corroborated. For example, documentation of apnea is frequently observer dependent and open to bias. When Bifano and co-workers performed objective measures of apnea using pneumocardiography, they could find no increased number or severity of apneas by hemoglobin range(30). There is a remarkable paucity of fully published randomized trial data, we are aware of only two performed in newborns. Of these one is historical and was completely un-powered for meaningful deductions (n=10 per arm) (31), and moreover both were conducted in infants with birthweight of little current relevance (32). Both also address transfusion levels more typical of convalescent babies. Of three more recent randomized trials addressing blood transfusion from birth to discharge(33,34,35), addressed short-term goals only, were of small size and moreover are as yet unpublished. Another recent trial of transfusion was only confined to transfusion after 29 days of life, was underpowered for short-term outcomes, and did not address follow-up(36). None of these trials have addressed, or have the power to address, long-term outcome. The starting point now in weighing the risks and benefits of transfusion in intensive care environments must be the pivotal trial of Hebert P et al(37). This trial although performed in adults undergoing ICU care, posed questions that are familiar to the neonatal intensivists. Namely the hypotheses being examined was whether or not, a higher hemoglobin by conveying a higher potential oxygen carrying capacity would confer benefit. In a startling result, the trials showed a higher mortality when a higher hemoglobin trigger was used as an indication to transfuse. What End Points Should Such a Trial Entail? The range of outcomes to measure must surely reflect long-term status(38), and go beyond simple neonatal mortality. About 53% of extremely low birth-weight infants <1000 g birthweight, will survive without apparent major impairment by 18 months of corrected age(39). However it is inappropriate to use this as the end point of trials now, when it is even unproven as to whether there is any interim short term benefit. This was not the case with prophylactic indomethacin that justifiably needed a primary objective at 18 months corrected age. But undoubtedly the primary objectives of current trials should include the main determinants of neonatal outcomes. A lot of these have been over the years increasingly linked to free radical disease. The effects of blood transfusion on iron balance have recently received attention. 15 ml of packed red cells contains about 200 mmol (11 mg) of iron, or about one fifth of the iron stores of a 1000 g infant(22,40). There is a direct correlation between serum ferritin levels and blood transfusion in infants(41). Infants receiving multiple transfusions show clear evidence of iron overload, as manifest by the presence of free iron and saturated transferrin levels in the plasma and histopathological evidence of iron overload in the liver(42,45). The presence of this powerful oxidant may contribute to the development or exacerbation of chronic lung disease (CLD) or retinopathy of prematurity (ROP) (43-48). On the other hand, extreme lowbirth weight (ELBW) infants have substantial iron losses associated with phlebotomy and limited internal intake, and blood transfusion has long been regarded as the main effective source of iron nutrition for infants receiving prolonged intravenous alimentation. The short-term consequences of maintaining ELBW infants at relatively high versus relatively low hemoglobin levels may include differences in mortality and in the neonatal occurrence of the acute morbidities of ROP, CLD, and white matter injury (periventricular leukomalacia and/or ventriculomegaly). Funding: None. Competing interests: The corresponding author is the Principal Investigator (with Dr.R.Whyte) of a multi-national multi—centre trial, funded by Canadian Institute of Health Research, called (Premature In Need of Transfusions(PINT)) which is estimated to complete enrollment in 2003. Nandkishor S Kabra, Clinical Fellow Haresh Kirpalani, Professor of Pediatrics Address for Correspondence: Professor Haresh Kirpalani Department of Pediatrics Room 3N27A McMaster University 1200 Main Street West Hamilton, Ontario L8N 3Z5 Phone: (905) 521-2100 ext.73024 Email: [email protected] |
|
References |
|
