|
|
| |||||||||||||||||||||||||||||||||
Background Early administration of high doses of dexamethasone may reduce the risk of chronic lung disease in premature infants but can cause complications. Whether moderate doses would be as effective but safer is not known.
Methods We randomly assigned 220 infants with a birth weight of 501 to 1000 g who were treated with mechanical ventilation within 12 hours after birth to receive dexamethasone or placebo with either routine ventilatory support or permissive hypercapnia. The dexamethasone was administered within 24 hours after birth at a dose of 0.15 mg per kilogram of body weight per day for three days, followed by a tapering of the dose over a period of seven days. The primary outcome was death or chronic lung disease at 36 weeks' postmenstrual age.
Results The relative risk of death or chronic lung disease in the dexamethasone-treated infants, as compared with those who received placebo, was 0.9 (95 percent confidence interval, 0.8 to 1.1). Since the effect of dexamethasone treatment did not vary according to the ventilatory approach, the two dexamethasone groups and the two placebo groups were combined. The infants in the dexamethasone group were less likely than those in the placebo group to be receiving oxygen supplementation 28 days after birth (P=0.004) or open-label dexamethasone (P=0.01), were more likely to have hypertension (P<0.001), and were more likely to be receiving insulin treatment for hyperglycemia (P=0.02). During the first 14 days, spontaneous gastrointestinal perforation occurred in a larger proportion of infants in the dexamethasone group (13 percent, vs. 4 percent in the placebo group; P=0.02). The dexamethasone-treated infants had a lower weight (P=0.02) and a smaller head circumference (P=0.04) at 36 weeks' postmenstrual age.
Conclusions In preterm infants, early administration of dexamethasone at a moderate dose has no effect on death or chronic lung disease and is associated with gastrointestinal perforation and decreased growth.
Data from some, though not all, clinical trials suggest that the early administration of dexamethasone (with the first dose given within 24 to 48 hours after delivery) may reduce the risk of chronic lung disease.6,7,8,9,10,11 In these studies, the initial dose of dexamethasone was high (
0.5 mg per kilogram of body weight per day), and many infants had adverse effects, such as hypertension or hyperglycemia. We performed a study to determine whether treatment with a moderate dose of dexamethasone would reduce the risk of chronic lung disease and have minimal adverse effects.
Methods
Infants
Inclusion criteria were a birth weight of 501 to 1000 g, treatment with mechanical ventilation within 12 hours after birth, and the presence of an indwelling vascular catheter. For infants whose birth weight was 751 to 1000 g, additional inclusion criteria were ventilation at a fraction of inspired oxygen of 0.3 or more and the administration of at least one dose of surfactant. We excluded infants with major congenital anomalies, congenital nonbacterial infection, findings indicating a very low likelihood of recovery (pH <6.8 or hypoxemia with bradycardia for more than two hours), or prior postnatal treatment with a glucocorticoid. The study was conducted at 13 participating centers between February 1998 and February 1999. The protocol was approved by the institutional review board at each center, and written informed consent was obtained from a parent of each infant.
Randomization
In a two-by-two factorial design, we tested both dexamethasone treatment and a strategy of minimal ventilatory support (permissive hypercapnia). Infants were randomly assigned to one of four groups according to the study medication (dexamethasone or placebo) and ventilatory treatment (routine treatment with the goal of maintaining the partial pressure of carbon dioxide at a level below 48 mm Hg or minimal ventilatory support with the goal of maintaining the partial pressure of carbon dioxide at a level above 52 mm Hg)12 with the use of a random, permuted-block algorithm. The treatment assignments were stratified according to the center and the infant's birth weight (501 to 750 g or 751 to 1000 g). All staff members except the pharmacist were unaware of the drug-group assignments, but the ventilatory-group assignments were not masked.
Study Protocol
Treatment with the study medication was initiated within 24 hours after birth. The dexamethasone-treated infants received a 10-day tapered course (0.15 mg of dexamethasone per kilogram per day for three days, followed by 0.10 mg per kilogram for three days, 0.05 mg per kilogram for two days, and 0.02 mg per kilogram for two days), with the daily dose divided in half and given at 12-hour intervals intravenously or orally, if an intravenous catheter was no longer in place. The initial dose was approximately equivalent to five times the estimated cortisol-replacement dose.13 The infants in the placebo groups received equal volumes of saline. During the 10-day treatment period, we discouraged the prescription of open-label glucocorticoids by the attending neonatologist, and we recorded any use of glucocorticoid therapy during hospitalization.
Outcomes
The primary outcome was the combination of death by 36 weeks' postmenstrual age or chronic lung disease (defined by a need for supplemental oxygen at least 12 hours per day) at 36 weeks' postmenstrual age. Secondary outcomes included chronic lung disease, death by 36 weeks' postmenstrual age, a need for supplemental oxygen 28 days after birth, open-label glucocorticoid treatment, the level of respiratory support (mechanical ventilation, continuous positive airway pressure, or supplemental oxygen alone) at 28 days after birth and at 36 weeks' postmenstrual age, and the duration of oxygen therapy, ventilatory support, and the hospital stay.
During the 10-day intervention period, we recorded hypertension (systolic pressure, >80 mm Hg), drug treatment for hypertension, hyperglycemia (blood glucose concentration, >180 mg per deciliter [10 mmol per liter]), insulin treatment for hyperglycemia, and evidence of upper gastrointestinal bleeding (a heme-positive gastric aspirate or emesis). We also recorded nosocomial infection, necrotizing enterocolitis, spontaneous gastrointestinal perforation, pulmonary interstitial emphysema, pneumothorax, pulmonary hemorrhage, patent ductus arteriosus, intracranial hemorrhage, periventricular leukomalacia, retinopathy of prematurity, and growth at the time of discharge or death or at 120 days of age, if the infant remained hospitalized. Research nurses collected all study data according to defined criteria and transmitted the data to a central coordinating center.
Statistical Analysis
Using the Neonatal Research Network data base, we calculated that to determine whether treatment with dexamethasone would reduce the primary outcome from 55 percent to 44 percent (i.e., reduce the relative risk of the outcome by 20 percent), we would need a sample of 532 infants in each group. To ensure an adequate number of infants to evaluate the neurodevelopmental outcome, which we expected to do at 18 months' corrected age, we planned to enroll 600 infants in each group.
We performed an intention-to-treat analysis. Base-line data for infants enrolled in the study and for eligible infants who were not enrolled were compared by t-tests for continuous variables and by chi-square tests for categorical data. Logistic regression was used to analyze differences in outcomes and complications between the treatment groups. Multiple logistic-regression analysis was used for categorical variables with more than two values (e.g., respiratory support). Initially, the analyses included dexamethasone treatment, ventilatory treatment, and an interaction term for dexamethasone and ventilatory treatment as factors. Because none of the interactions were significant, the analyses were repeated without the interaction term, and we report the resulting P values for the main effects.
Results
The trial was monitored by an independent data and safety monitoring committee. The committee's initial evaluation, performed because of a high rate of unanticipated adverse events, identified frequent gastrointestinal perforations among the infants treated with dexamethasone. Because of the uncertainty involved in weighing the relative importance of potential benefits and adverse outcomes, the committee recommended continuation of the trial with a modification of the consent form to include this complication. However, the steering committee voted to terminate the trial.
Infants
During the study period, 340 infants were eligible for enrollment, and 220 were enrolled. The other 120 eligible infants were not enrolled because of a parent's refusal (55 percent), the unavailability of a parent to provide consent or failure to seek consent (41 percent), the physician's refusal (2 percent), or other, unknown reasons (2 percent). The infants who were not enrolled were similar to the enrolled infants with regard to birth weight (mean, 743 g in the group of unenrolled infants and 735 g in the enrolled group), gestational age (mean, 25.7 and 25.6 weeks, respectively), male sex (51 percent and 52 percent, respectively), vaginal delivery (42 percent in both groups), and antenatal glucocorticoid therapy (77 percent and 75 percent, respectively), but they differed in racial distribution (26 percent white vs. 41 percent, 52 percent black vs. 47 percent, and 20 percent Hispanic vs. 10 percent, respectively).
Because the effect of dexamethasone treatment did not vary according to the type of ventilatory treatment, the ventilatory-treatment groups were combined for the purpose of analysis. The base-line characteristics of the infants in the dexamethasone and placebo groups were similar (Table 1).
|
The relative risk of death or chronic lung disease at 36 weeks' postmenstrual age in the dexamethasone group was 0.9 (95 percent confidence interval, 0.8 to 1.1) (Table 2). The relative risk did not differ significantly between the two birth-weight groups. Mortality at 36 weeks' postmenstrual age and the rate of chronic lung disease among the infants who survived also did not differ significantly between the dexamethasone and placebo groups.
|
The mean (±SD) proportion of doses of the study drug that were actually given was lower in the dexamethasone group (93±16 percent) than in the placebo group (98±12 percent, P=0.01), reflecting in part doses withheld because of complications potentially attributable to the study drug. Infants in the dexamethasone group were less likely than those in the placebo group to receive open-label glucocorticoid treatment during hospitalization (34 percent vs. 51 percent, P=0.01). Only one infant in the dexamethasone group received open-label glucocorticoid treatment during the 10-day intervention period, as compared with eight infants in the placebo group. Among the infants who received supplemental treatment with open-label dexamethasone, the mean duration of treatment was 25±33 days in the dexamethasone group and 27±35 days in the placebo group (P=0.82). The proportion of infants who required mechanical ventilation, continuous positive airway pressure, or supplemental oxygen alone 28 days after birth or at 36 weeks' postmenstrual age did not differ significantly between the two study groups. Similarly, there were no significant differences between the groups in the duration of oxygen therapy or mechanical ventilation or in the median hospital stay among either infants who survived or those who did not.
Pulmonary interstitial emphysema was diagnosed less frequently in the dexamethasone group than in the placebo group (relative risk, 0.4), although the frequencies of pneumothorax, pulmonary hemorrhage, and patent ductus arteriosus did not differ significantly between the two groups (Table 3). The rates of other outcomes ascertained at death or discharge or at 120 days among hospitalized infants did not differ significantly between the dexamethasone and placebo groups (Table 3).
|
A higher proportion of infants in the dexamethasone group than in the placebo group had hypertension or received antihypertensive drugs (Table 4). Although the frequency of hyperglycemia was similar in the two groups, a larger proportion of infants in the dexamethasone group were treated with insulin. Upper gastrointestinal bleeding was uncommon, and the proportion of infants with this complication did not differ significantly between the two groups.
|
Perforation appeared to be associated with indomethacin treatment within the first 24 hours (P= 0.02), and the effect of dexamethasone on the perforation rate also appeared to be greater in the presence of indomethacin than in its absence (Table 5). Perforation occurred in 19 percent of infants treated with both dexamethasone and indomethacin, in 2 percent of those treated with dexamethasone alone, in 5 percent of those who received placebo and indomethacin, and in none of the infants who received only placebo (Table 5). Although indomethacin treatment was not randomly assigned, the difference in the rates of perforation between the infants who received indomethacin and those who did not was significant in both the dexamethasone group and the placebo group (P=0.05).
|
|
Unlike some previous investigators,6,7,8,9 we found no significant difference in the relative risks of chronic lung disease at 36 weeks' postmenstrual age, death, or the combined outcome in extremely-low-birth-weight infants treated with dexamethasone or placebo. However, since a criterion for enrollment in our study was a birth weight of 501 to 1000 g, our infants were relatively immature and at high risk for a poor respiratory outcome.
Infants in the dexamethasone group were less likely than those in the placebo group to require oxygen 28 days after birth, a finding that may be related to an antiinflammatory effect of dexamethasone treatment.14,15 Furthermore, infants in the placebo group were more likely than those in the dexamethasone group to be treated with open-label dexamethasone. Since the decision to administer open-label dexamethasone was made by the attending neonatologist, these infants may have had a poorer clinical status than those who received early treatment with dexamethasone. An increased use of subsequent glucocorticoid treatment in the placebo group has been noted in other trials of early systemic9,10 or inhaled16 glucocorticoid treatment, and such use may minimize differences in the respiratory outcome between the glucocorticoid and placebo groups.
Although the risk of necrotizing enterocolitis did not differ between the two study groups, the rate of spontaneous gastrointestinal perforation within the first two weeks in the dexamethasone group was more than three times that in the placebo group, and this complication appeared to be associated with the administration of indomethacin. Because indomethacin treatment was not assigned randomly, the infants who received indomethacin may have been more susceptible to perforation. Nevertheless, the high perforation rate in the dexamethasone group was unanticipated and resulted in termination of the trial.
Spontaneous perforation has been reported in very-low-birth-weight infants17,18 and has also been associated with dexamethasone treatment for chronic lung disease19 and indomethacin treatment for patent ductus arteriosus.20,21,22 The small numbers of extremely-low-birth-weight infants enrolled in previous trials of dexamethasone may have limited the ability to detect this adverse event. In a recent large trial of a short course of dexamethasone given soon after birth, perforation during the first week occurred in 8 percent of the dexamethasone-treated infants and in 1 percent of the infants who received placebo, although there was no significant difference between the groups in the overall rate of perforation.9 Similarly, in a large trial of a 12-day course of dexamethasone or placebo, perforation occurred more often in the dexamethasone group, although the difference was not statistically significant.10
The mechanism of perforation may be related to the role of prostaglandins in maintaining gastrointestinal mucosal integrity.23 Glucocorticoids and indomethacin inhibit prostaglandin production at two points in the synthetic pathway,24,25 perhaps explaining the association with perforation.
Hypertension and hyperglycemia are recognized complications of dexamethasone therapy.7,9,10,26,27 In our study, hypertension and insulin treatment were more frequent in the dexamethasone group than in the placebo group, although the rates of hypertension and insulin treatment were lower in our dexamethasone-treated infants than in similar infants given a higher dose of dexamethasone and a longer course of treatment.10 Use of other glucocorticoids or physiologic replacement28 rather than therapeutic doses may further reduce complications.
Dexamethasone treatment has been reported to have both transient and sustained negative effects on growth.7,10,29 In our study, the dexamethasone-treated infants weighed less than the placebo-treated infants at the end of the intervention period. In addition, the infants who received dexamethasone weighed less and had a smaller head circumference at 36 weeks' postmenstrual age, even though a larger proportion of infants in the placebo group were subsequently treated with open-label dexamethasone. Extremely-low-birth-weight infants may be especially susceptible to the catabolic effects of glucocorticoid treatment30 during the early postnatal period, when they are likely to receive too few calories, and this susceptibility may affect their subsequent growth.31
In summary, we found that a 10-day tapered course of dexamethasone given at a moderate dose had no discernible effect on chronic lung disease or mortality in extremely-low-birth-weight infants. The dose we used, although substantially lower than the initial doses used in other trials or in clinical practice, was associated with an increased risk of spontaneous gastrointestinal perforation, as well as with known complications of glucocorticoid therapy. The risk of perforation appears to be associated with concomitant indomethacin treatment. Given these serious complications and the lack of a discernible benefit, we believe that early treatment with dexamethasone to prevent chronic lung disease in extremely-low-birth-weight infants is not indicated.
Supported by cooperative agreements with the National Institute of Child Health and Human Development (U10 HD34167, U10 HD34216, U10 HD21373, U10 HD27881, U10 HD21385, U10 HD27853, U10 HD27904, U01 HD21397, U01 HD36790, U10 HD27851, U10 HD21364, U10 HD27871, and U10 HD21415) and by grants from the General Clinical Research Centers Program (M01 RR 02635, M01 RR 02172, M01 RR 00997, M01 RR 08084, M01 RR 06022, M01 RR 08084, and M01 RR 00070).
We are indebted to Drs. Gordon Avery, Mary D'Alton, John Fletcher, Christine Gleason, Maureen Maguire, Carol Redmond, and Robin Roberts for their contributions as members of the data and safety monitoring committee; to Drs. John Sinclair and Mark Klebanoff for their helpful review of the manuscript; and to our medical and nursing colleagues and the infants and their parents who participated in the study.
Source Information
From Brigham and Women's Hospital, Boston (A.R.S.); the University of Alabama at Birmingham, Birmingham (W.A.C.); the University of Texas at Houston, Houston (J.E.T.); the University of New Mexico, Albuquerque (L.-A.P.); the National Institute of Child Health and Human Development, Bethesda, Md. (L.L.W.); Wayne State University, Detroit (S. Shankaran); the University of Cincinnati, Cincinnati (E.F.D.); Women and Infant's Hospital, Providence, R.I. (W.O.); the University of Miami, Miami (C.R.B.); Research Triangle Institute, Research Triangle Park, N.C. (S. Saha, W.K.P.); and Emory University, Atlanta (B.J.S.). Other authors were Avroy A. Fanaroff, M.B., B.Ch., Case Western Reserve University, Cleveland; Richard A. Ehrenkranz, M.D., Yale University, New Haven, Conn.; Sheldon B. Korones, M.D., University of Tennessee at Memphis, Memphis; and David K. Stevenson, M.D., Stanford University, Stanford, Calif.
Address reprint requests to Dr. Stark at Newborn Medicine, CWN-6, Brigham and Women's Hospital, 75 Francis St., Boston, MA 02115, or at astark{at}uptodate.com.
References
Other members of the National Institute of Child Health and Human Development Neonatal Research Network who participated in the study were as follows: University of Alabama at Birmingham — M.V. Collins; Brigham and Women's Hospital — K.A. Fournier; Case Western Reserve University — M. Hack, N. Newman; University of Cincinnati — A. Jobe (chair of steering committee), M. Mersmann; Emory University — E. Hale; University of Miami — S. Duara, A.M. Worth; National Institute of Child Health and Human Development — S.J. Yaffe, E.M. McClure; University of New Mexico — C. Backstrom; Research Triangle Institute — B. Hastings; Stanford University — M.B. Ball; University of Tennessee at Memphis — H. Bada, T. Hudson; University of Texas Southwestern Medical Center — A. Laptook, S. Madison; Wayne State University — G. Konduri, G. Muran; Women and Infants' Hospital — B. Stonestreet, A. Hensman; Yale University — P. Gettner.
| |||||||||||||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
