 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
To the Editor.—
Bell et al1 suggest that restrictive transfusion policies for preterm infants may result in an increased incidence of adverse neurology, either a grade 4 intraventricular hemorrhage (IVH) or periventricular leukomalacia. However, we are concerned that limitations of the trial design have led to an overinterpretation of the data.
The trial specifically excluded infants with a significant patent arterial duct. Patent arterial duct is an extremely common finding in this population yet was not defined robustly. Was echocardiography routinely used, and if so on which variables was significance decided? This may have a crucial effect on the generalizability of the study. Failure to stratify for gender (a factor known to be associated with developmental outcome) led to an excess of males in the restrictive group, although a balanced design could have attempted to control for this.2 However, a more important concern with stratification was the use of weight strata that meant gestation could not be similarly controlled. This is unfortunate, because one of the principal outcomes, an abnormality detected on "late" cranial ultrasound, was only performed based on gestational criteria. Half of the study cohort did not receive a late ultrasound scan.
Power was based on the predicted difference in number of transfusions between the 2 groups. Secondary outcomes need to be interpreted with caution. The authors assumed that infants not receiving a late ultrasound did not have significant adverse neurology. The use of interpolation (eg, assuming that 
1 infants not receiving a late scan actually had significant changes) would have enabled the reader to make a more balanced assessment of whether the purported risk associated with restrictive criteria may have occurred by chance.
It is not clear why infants were not enrolled or approached at a standardized time, but we know that some infants may have received 2 transfusions before participation. Infants received their first transfusion at a median postnatal age of 3 days in the liberal-transfusion and 8 days in the restrictive-transfusion group, but infants in both groups may have received transfusions on day 1. The principal difference between the groups was in the incidence of grade 4 IVH, but we know that this is likely to occur in the first 2 to 3 days in 
80% of infants.3,4 The authors were not able to perform an ultrasound before enrollment, so we cannot know the timing for certain. If the neurologic difference between the groups is a consequence of the transfusion criteria, are we correct in interpreting the authors' postulate that late transfusions (ie, after the first 2–3 days) are responsible for the excess of grade 4 IVH?
It is difficult to believe that there is a genuine difference in IVH likely to affect long-term outcome (ie, at least grade 3 IVH) between the groups. Eight infants in the liberal-transfusion group but only 1 in the restrictive-transfusion group were diagnosed with a grade 3 IVH, compared with 0 and 4 infants, respectively, with a grade 4 IVH. If the etiologic mechanisms for both grades 3 and 4 are the same, it is surprising that there was 8 times less grade 3 IVH in the restrictive-transfusion group unless, as with the excess of grade 4 IVH in that group, the results occurred by chance.
The authors are to be commended for tackling an important issue in neonatology. They have confirmed that transfusion rates and donor exposures can be minimized by careful practice. However, we do not feel the study methodology adequately identified a risk of restrictive transfusion policies.
REFERENCES
Related articles in Pediatrics:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2006 American Academy of Pediatrics. All rights reserved. |
||
|
||
