 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
To the Editor.—
We read with interest the recent joint statement of the AAP, Committee on the Fetus and Newborn and the Canadian Paediatric Society, Fetus and Newborn Committee,1 the recent letter to the editor by Taylor et al,2 and the response by Dr Barrington3 on the topic of postnatal glucocorticoids for chronic lung disease in the neonate. The use of steroids, dexamethasone in particular, to treat or prevent chronic lung disease (CLD) is controversial. Although we applaud the efforts of the Academy to bring some sense to this quagmire, we have serious reservations about some of the conclusions and recommendations.
Our major concern is that conclusions in the AAP report relating to neurodevelopmental outcome after glucocorticoid treatment were based largely on 2 published meta-analyses performed on 54 and 8 studies5 that appear to be seriously flawed. In a comment on the latter paper, Taylor et al2 discussed the impropriety of combining 8 heterogeneous studies that differed in age at initiation of glucocorticoid therapy, dose and duration of therapy, and tests (and timing of these tests) used to determine neurodevelopmental outcome. Their succinct and valid arguments apply equally well to the meta-analysis of 5 studies,4 and need not be repeated. It was disappointing to note that the AAP report accepted the findings of these 2 analyses without any stated reservations and then went on to recommend restrictive conditions for using postnatal glucocorticoids.
Review of the 11 studies listed in the AAP report reveals that in 5, glucocorticoid therapy was initiated in the first 4 days after birth and in the remaining 6, therapy was begun at 7 days’ postnatal age or later. Another way to look at the published data is to note that in 3 of the 5 studies that involved initiation of therapy within 4 days after birth there was at least one significant neurodevelopmental defect at follow-up. However, of the 6 studies in which therapy was started at 1 week or later, 1 reported a significant adverse neurodevelopmental finding, but 5 did not. The study by O’Shea et al,6 which reported an increased incidence of cerebral palsy (but no difference in the Mental Development Index or the Psychomotor Development Index, used an initial dose of 0.5 mg/kg/day dexamethasone and continued administration for a total of 42 days.
In fact, apart from the study of Stark et al7 that used an initial dose of 0.15 mg/kg/day, all the published trials used an initial dose of dexamethasone of 0.5 mg/kg/day or higher. There is no evidence that such a high dose is necessary. The recent study of Durand et al8 demonstrated improved pulmonary compliance in infants with CLD using an initial dose of 0.2 mg/kg/day.
There clearly should be concern relating to the prophylactic use of dexamethasone in the first few days after birth in an attempt to prevent CLD, and to the use of 28- to 42-day courses of dexamethasone for treatment of CLD. The primary goal of glucocorticoid therapy should be to facilitate extubation—one of the few documented benefits of glucocorticoid administration for CLD. Hence, if glucocorticoids are used, it seems rational to limit therapy to a short period such as 5 days. If extubation is not accomplished in this time, there is probably little to be gained and much to be lost be a prolonged course.
We agree with the recommendation for carefully designed, randomized, controlled trials of dexamethasone (or other glucocorticoids such as hydrocortisone) for treatment of CLD. However, such trials should include appropriate measures of efficacy along with follow-up to detect adverse consequences including neurodevelopmental delay.
Herein lies a problem. How does one define "appropriate" measures of efficacy? Altered mortality is an easily definable outcome, but it is a very crude measure of efficacy. Even measuring the incidence of CLD is highly suspect because this diagnosis does not distinguish between infants with vastly different degrees of illness. We believe that carefully designed randomized studies, without contamination of the control group, which measure pulmonary function, exercise tolerance, growth and neurodevelopmental outcome will help us define the appropriate circumstances, timing and dose of dexamethasone for CLD.
What do we do in the meantime? We do not agree with the recommendation that outside the context of randomized, controlled trials, current use of dexamethasone should be limited to infants on maximal ventilator and oxygen support as suggested in the AAP report. We believe that because dexamethasone, when given between 7 and 14 days after birth, does decrease the incidence of CLD and mortality9 withholding therapy until an infant is on maximal support is too late. Furthermore, we have found that the assertion by Dr Barrington3 that, if appropriately informed, most parents would not agree to the use of dexamethasone is simply incorrect. Obviously parents will respond to the information they are given and this in turn will depend on how one interprets the literature. Hopefully this will be done in a thoughtful and objective fashion. Until the data from the much needed trials are available, one must be prudent and cautious in the use of dexamethasone. If it is used, the shortest course, probably no more than 5 days, and the lowest doses possible, perhaps 0.15–0.2 mg/kg/day initially, should be used with parental consent.
REFERENCES
In Reply.—
Dr Jacobs and his colleagues raise some important points in the ongoing debate regarding the issue of postnatal corticosteroids in preterm infants. They also point out some of the limitations in the published meta-analyses of neurodevelopmental outcome.1,2 However, they are incorrect in stating that the Committee on Fetus and Newborn of the American Academy of Pediatrics and the Fetus and Newborn Committee of the Canadian Paediatric Society accepted these meta-analyses without reservation.3 In fact, the systemic review was completely redone, with the inclusion of 3 additional follow-up studies that increased the sample size to 870 infants evaluated at 1 year of age or later. Because the Committees also had concerns regarding the heterogeneity of the study designs, the data were not combined in a meta-analysis, but were listed in tabular form. The recommendations were therefore not based on the results of the "flawed meta-analyses" but on the overall weight of the concerns regarding consistent adverse effects of postnatal corticosteroids on neurodevelopment.
Are the data regarding harm absolutely secure? Clearly no. It is possible that some other effect is causing the apparent relationship between steroids and neurodevelopmental impairment. However, this relationship is biologically plausible, consistent with animal data4 and supported by observational studies.5 Furthermore, the randomized trials with low contamination6–8 all suggest that postnatal administration of dexamethasone causes substantial adverse central nervous system effects.
No long-term benefit of postnatal corticosteroids has yet been proven, despite randomized trials that have included many hundreds of infants. It is possible that a lower dose and a shorter course of dexamethasone, or a corticosteroid other than dexamethasone, may demonstrate currently unproven pulmonary benefits, sufficient to outweigh the concerns regarding long-term neurodevelopmental outcome. This remains to be proven by appropriately designed trials that are adequately powered to test long-term effects, as recommended by the statement. However, it seems unlikely that additional trials using already tested regimens will yield previously undocumented benefits. Our efforts may be better directed toward investigating other strategies of lung protection.
The short-term benefits of postnatal corticosteroids include improved lung mechanics and gas exchange leading to a reduction in ventilator duration. This improvement leads to a decrease in the need for oxygen at 36 weeks’ postmenstrual age, and thus a reduced frequency of the diagnosis of chronic lung disease, but does not appear to reduce the numbers of children needing home oxygen therapy. Furthermore, the short-term benefits are associated with substantial risks, including gastrointestinal bleeding and perforation, growth impairment, infection, hypertension, hyperglycemia, and cardiac hypertrophy.
Dr Jacobs and colleagues suggest that dexamethasone should be given at 1 to 2 weeks of age to assist in extubation, and recommend an untested treatment regimen. We are aware of no data showing that infants who are ready to be extubated are more likely to be successfully extubated if they receive steroids than if they do not. It is incumbent on those who would propose such a course of action to prove that it is effective and safe.
A primary guiding principle of medicine is "primum non nocere." By universal use of antenatal steroids whenever possible, early surfactant, limiting tidal volumes, having reasonable ventilation goals rather than normalizing PCO2, limiting oxygen administration to the least necessary, assuring adequate vitamin A status, avoidance of fluid boluses and judicious fluid administration in early postnatal life, lung injury can be minimized. This can be accomplished without postnatal corticosteroids.
REFERENCES
This article has been cited by other articles:
 |
|
 |
|
  L. B. Clerch, A. S. Baras, G. D. Massaro, E. P. Hoffman, and D. Massaro DNA microarray analysis of neonatal mouse lung connects regulation of KDR with dexamethasone-induced inhibition of alveolar formation Am J Physiol Lung Cell Mol Physiol, February 1, 2004; 286(2): L411 - L419. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Perlman Concern About Fetus and Newborn Committee Statement on Corticosteroid Use Pediatrics, November 1, 2002; 110(5): 1034 - 1034. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2002 American Academy of Pediatrics. All rights reserved. |
||
|
||
