Randomized, Controlled Trial of Dexamethasone in Neonatal Chronic Lung Disease: 13- to 17-Year Follow-up Study: I. Neurologic, Psychological
http://www.100md.com
《小儿科》
the National Perinatal Epidemiology Unit, Oxford University, Oxford, United Kingdom
ABSTRACT
Objectives. To study neurologic, educational, and psychological status in adolescence of neonates enrolled in a double-blind, randomized, controlled trial of dexamethasone therapy for chronic lung disease.
Participants. A total of 287 infants who were chronically dependent on supplementary oxygen and were 2 to 12 weeks of age were recruited from 31 centers in 6 countries to a randomized, controlled trial of dexamethasone base (0.5 mg/kg per day for 1 week); 95% of survivors were reviewed at 3 years. Survivors from the 25 British and Irish centers were retraced at 13 to 17 years of age.
Outcome Measures. Nonverbal reasoning, British Picture Vocabulary Scale, Goodman Strengths and Difficulties Questionnaire behavior scores, school national test results, teacher ability ratings, and parental and general practitioner questionnaires.
Results. A total of 195 children were eligible for the follow-up study. Information was available for 150 children (77%), with 142 (73%) being assessed in home visits. No baseline differences were detected between the children included in the follow-up study and those not included. There was a slight excess of cerebral palsy in the steroid group, which was not statistically significant (relative risk: 1.58; 95% confidence interval: 0.81–3.07). Overall disability rates in both groups were high (21% moderate and 14% severe), but with no difference between the 2 groups (for severe disability, relative risk: 0.84; 95% confidence interval: 0.37–1.86).
Conclusions. Information was obtained for 150 adolescents randomized to receive dexamethasone or placebo for neonatal chronic lung disease. Rates of disabilities and educational difficulties were high, but with no significant differences between the 2 groups. Some use of open-label steroids in the placebo group plus losses to long-term follow-up monitoring reduced the power of this study to detect clinically important differences, and this study cannot rule out a real increase in cerebral palsy, as reported by others.
Key Words: neonatal chronic lung disease dexamethasone therapy randomized controlled trial developmental follow-up school performance
Abbreviations: CLD, chronic lung disease CP, cerebral palsy GP, general practitioner RR, relative risk CI, confidence interval
Neonatal chronic lung disease (CLD) affects 25% of very low birth weight infants,1 with an increased risk of adverse neurodevelopmental outcomes.2,3 The beneficial effects of prenatal corticosteroid therapy on fetal lung maturation were first shown >30 years ago,4 but a controlled trial of postnatal hydrocortisone therapy for infants with respiratory distress syndrome that was published the same year showed no benefit.5 Additional studies of prenatal corticosteroid therapy have confirmed the reduction in surfactant-deficient respiratory distress syndrome and neonatal mortality rates and secured the place of such therapy in obstetric practice,6 but the place of postnatal steroid treatment remains controversial. A multicenter, randomized, controlled trial of dexamethasone enrolled 287 infants 4 weeks of age with established CLD and showed a significant reduction in the duration of assisted ventilation but no reduction in mortality rates or overall hospital stay.7 Follow-up evaluation of survivors at 3 years showed no persisting benefits but no evidence of increased adverse outcomes,8 and it is these children who form the basis of the current report. A plethora of neonatal trials have resulted in 3 major systematic reviews of corticosteroid treatment, given "early" (age of 0–4 days, 21 trials),9 "moderately early" (age of 7–14 days, 7 trials),10 or "delayed" (age of >3 weeks, 9 trials),11 confirming earlier extubation and a reduction in CLD. Moderately early treatment was also associated with a significant reduction in mortality rates.10 Short-term side effects were common, but most seemed transient.
However, in 1998, a 2-year follow-up study of infants given prophylactic dexamethasone showed an increase in adverse neuromotor outcomes,12 sounding a note of caution. Follow-up data are available for less than one half of the trials included in systematic reviews but have produced conflicting results regarding a possible link with later cerebral palsy (CP). The early group of trials showed an increase in CP rates but no increase in rates of overall disability or combined adverse outcomes of death or CP,9 and the delayed group also showed a nonsignificant increase in CP rates.11 Conversely, moderately early treatment showed a trend for decreased CP rates, but with a smaller number of children being monitored.10 Prenatal steroid use was associated with a significant reduction in neonatal intraventricular hemorrhage rates6 and also a trend toward reduction in CP rates. Individual authors and professional bodies have indicated caution regarding neonatal steroid use and have highlighted the lack of published information on progress into later childhood.13,14 It was these concerns that led to a decision to perform an additional follow-up study in adolescence. The aim was to assess whether early exposure to dexamethasone had any long-term consequences with respect to neurodevelopmental and educational progress, as reported in this article, or growth, lung function, and general health, as reported in the accompanying article.
METHODS
Participants
Between 1984 and 1989, 287 infants with neonatal CLD were recruited to an international, randomized, placebo-controlled trial of dexamethasone therapy.7 Eligible infants were those with no major congenital malformations who were dependent on supplementary oxygen and whose condition was static or deteriorating at 2 to 12 weeks of age (median age: 4 weeks). Randomization was through telephone contact with a central office, with infants stratified according to center and whether they were still receiving assisted ventilation (61%) or not (39%). Treatment was with either dexamethasone phosphate (0.6 mg/kg per day for 1 week, equivalent to 0.5 mg/kg dexamethasone base) or matching saline placebo, given intravenously or orally and with the option of a second tapering 9-day course. Clinicians were allowed to prescribe open-label dexamethasone if there was life-threatening deterioration, but they remained blinded to the trial treatment allocation. The trial was conducted in 31 centers in 6 countries, and a 3-year follow-up study was undertaken.8 Children from continental Europe and North America were excluded from additional follow-up monitoring because of the lack of a reliable system for tracing them. Of children recruited from British and Irish centers, 99% had been successfully evaluated at 3 years of age. No attempt had been made to keep in contact with the families in the intervening years.
Research Ethics Committee Approval
The original, randomized, controlled trial was approved by the research ethics committee at each of the participating centers, and informed parental consent was obtained. Approval for the present study was obtained from the Anglia and Oxford Multicenter Research Ethics Committee, and the appropriate local research ethics committees were informed. The study complied with the ethical guidelines of the British Educational Research Association. Informed consent was obtained from the children and their parents for each aspect of the study. Families have been offered a summary of the overall findings of the study.
Tracing and Contact
Initial contact was made with the last known general practitioners (GPs), to check whether the children were still registered with them. A search for the remainder of the children was made through the National Health Service Central Register, to ascertain any deaths, and the Family Health Services Authority, to determine the current GPs for survivors. A letter was sent requesting the GPs to forward details of the study to the families. A reply slip and prepaid envelope to the study office were enclosed. If parents failed to respond, then a second mailing was sent. For continued nonresponders, GPs were approached and asked whether there was any known reason for nonresponse; they were asked to flag the patients' notes and to speak to them about the study if they attended the surgery. No direct contact was made with families without their agreement.
Consideration was given to obtaining consent from the children themselves, who were then adolescents. A pilot study of 67 families ascertained that 58 children were already aware of their preterm birth and of the trial and no parent anticipated difficulties in talking to the child about it. In addition to the explanatory letter to parents, a separate information leaflet was enclosed for parents to give to the child. This leaflet was drafted with input from a group of healthy teenagers and also parents of formerly preterm infants. Consent was obtained from parents and children for each part of the study separately. Therefore, families could agree to a home visit but decline contact with their school. Family practitioners, with the families' consent, were given the results of nurses' assessments but not the teachers' reports. All home visits were conducted between September 2001 and December 2002.
Assessment Tools
All questionnaires and information leaflets are available from the author or electronically (www.npeu.ox.ac.uk/dex). The families completed a questionnaire on functional status, diagnoses of potentially disabling conditions (visual or hearing impairments, learning disabilities, CP, and epilepsy), and the child's schooling. Questions on respiratory symptoms are detailed in the accompanying article. GPs also completed a questionnaire, reporting on any known functional problems, diagnoses, and hospital admissions. No specific definitions were provided, and the diagnosis of CP was made by the pediatrician responsible for each child's care. Children were visited at home by 1 of 3 research nurses, who were blinded to the children's original treatment allocation. To assess different areas of cognitive ability, a nonverbal reasoning test15 and the British Picture Vocabulary Scale16 were administered. These tests, published by the National Foundation for Educational Research, are used widely for British schoolchildren and have both been restandardized recently. Training was provided by staff members from the National Foundation for Educational Research. Results are expressed as a performance age, which is compared with the child's actual age to yield a quotient. The tests use a multiple-choice type of approach, and the nonverbal reasoning test can be completed by a child with poor reading skills or one with poor coordination or CP, because no drawing or fine motor skills are involved. The tests do not need to be administered by a trained psychologist. These tests could not be administered to 2 blind children, and information regarding their cognitive abilities was obtained from their specialist teachers.
Disability Rating
For the purposes of an overall disability rating, the results of the nonverbal reasoning test and the British Picture Vocabulary Scale were averaged as a proxy for IQ. Moderate disability was defined as 1 or 2 of the following: IQ 2 to 3 SDs below the mean, ambulatory CP, hearing deficits corrected with aids, impaired vision, or behavior disorder with a major impact on schooling. Severe disability was any of the following: IQ >3 SDs below the mean, wheelchair-dependent CP, uncorrectable hearing loss, blind (perception of light only), or 3 moderate disabilities.
School Assessment
Permission was sought from parents and children to contact their schools for completion of a teacher assessment form, including the child's most recent national tests results (Key Stage 2 at age 11 or Key Stage 3 at age 14 for England and Wales or the equivalent examinations for Scotland and Ireland). Teachers were also asked to estimate the child's position within an average mixed-ability class for a range of subjects, on an analog scale of 1 to 30 (with 1 being the lowest). The same questionnaire was used successfully for another cohort of preterm infants evaluated as teenagers.17 The teacher assessment also incorporated a standardized behavior inventory, the Strengths and Difficulties Questionnaire described by Goodman.18 The questionnaire includes an "impact score" summarizing the effects of the child's difficulties on the individual child, peer relationships, the teacher, and the whole class. Teachers were asked to leave no details of the study in the child's school records when they returned the questionnaire.
Sample Size, Outcome Measures, and Statistical Methods
The sample size was predetermined by the size required for the original randomized trial minus any deaths and enrollments from centers outside Britain and Ireland. Primary outcome measures were the proportions of children >2 SD below the mean for nonverbal reasoning or British Picture Vocabulary Scale scores. A sample size of 195 yielded an 80% chance of showing an increase in incidence in the primary outcomes from 20% to 39% or a decrease to 6%. Secondary outcome measures included the proportions of children with CP, sensory impairments, or severe neurosensory disabilities, school national test results, teacher ratings, and results of the Strengths and Difficulties Questionnaire. Statistical analyses were conducted with SPSS version 11 software (SPSS, Chicago, IL), with results given as relative risks (RRs) or differences with 95% confidence intervals (CIs); the 2 test for trend was used when appropriate. All analyses were performed on an intention-to-treat basis. Centers were ranked according to their use of open-label steroids, and a subanalysis was performed on the main outcome measures comparing children from the 11 centers where no open-label treatment was used with children from centers with low use (1–65%) and those from centers with high use (66–100%).
RESULTS
Of the original 287 infants randomized, 195 were eligible for follow-up assessments (Fig 1). Five children were known to have spastic quadriplegia and severe learning difficulties, and their GPs or health visitors provided information at 3 years. None have died since then, but they were not contacted for this later follow-up evaluation because it was thought that they would be unable to complete any of the test items. These children are included in the results on disabilities. Of the 190 children eligible for contact, 3 children were known to have emigrated, 1 child was known to have gone into foster care at the 3-year follow-up evaluation, and 1 other child who was evaluated at 3 years could not be traced at all. Thirty-four families failed to respond despite reminder letters, and 6 declined to participate. Health questionnaires were completed by 145 (76%) of the 190 families, a home visit was completed for 142 (75%), and a teacher questionnaire was returned for 118 children (63%). Including the 5 severely disabled children, some information was available for 150 (77%) of 195 survivors. There was no evidence of any differences between the children included in the follow-up study and those not included (Table 1). There was also no evidence that the comparison of the dexamethasone and placebo groups differed between those included in the follow-up study and those not included (Table 1). In the neonatal trial, placebo-treated infants were more likely to receive open-label steroids than were infants in the dexamethasone group (11% vs 39%), which reflects the short-term benefits of steroids. However, there was no difference in open-label steroid usage between those evaluated in adolescence and those lost to follow-up monitoring.
The key neurosensory outcomes are shown in Table 2. Nonverbal reasoning scores and vocabulary scores showed no significant differences between the 2 groups either in the proportions of children >2 SDs below the mean or in the mean values themselves. Twenty-nine children had a diagnosis of CP (dexamethasone group: 17 of 71 subjects; placebo: 12 of 79 subjects; RR: 1.58; 95% CI: 0.81–3.07). There was no difference between the groups in overall disability, whether moderate or severe, in the use of therapists, or in the incidence of seizures or hydrocephalus.
Teachers completed the Goodman Strengths and Difficulties Questionnaire (Table 4). In an average population of children, 10% are expected to have abnormal scores and an additional 10% to have borderline scores.18 These children showed an increased proportion with abnormal scores but no difference between the 2 groups, in either total scores (22% abnormal, as shown in Table 4) or individual subscores (conduct problems, emotional symptoms, hyperactivity, and peer problems; data not shown). Teachers reported some difficulties in behavior, emotion, or concentration for more than one half of the children, with these difficulties significantly affecting the child and the rest of the class. However, teachers reported better than average scores for prosocial behavior.
Use of open-label steroids for 39% of the placebo-treated children and 11% of the dexamethasone-treated children had the potential to reduce any differences between the 2 groups. A secondary analysis was therefore performed for the primary outcome measures for the 2 groups, stratified according to centers with no, low, or high open-label steroid usage (Table 5). For children from centers using no open-label steroids, the RRs of low nonverbal reasoning test scores, low British Picture Vocabulary Scale scores, and CP were greater than those for children from centers with low or high steroid usage, but all CIs overlapped 0, as did the ratio of RRs in interaction tests, which suggests that all of these differences could have occurred by chance.
DISCUSSION
This is the largest follow-up study through to adolescence of preterm infants treated with neonatal corticosteroid therapy. The results confirmed the high risk of adverse outcomes among infants with neonatal CLD; 35% had moderate or severe disabilities, including 19% with a diagnosis of CP. Twenty-one percent attended a "special needs" facility and two thirds were achieving below expected levels for their age in national educational assessments. Despite their difficulties, most children were seen as cooperative and considerate members of their schools, with high ratings on the prosocial behavior scale. However, no clear differences were seen between the 2 groups for any outcome measure. This is in contrast to reports for younger children given early neonatal dexamethasone treatment, which showed a significantly higher incidence of CP9 and a finding of lower IQ at 8 years of age.20 There are various possible explanations for these differences.
The first concern involves tracing of the children. At the 3-year follow-up assessment, only 1% of survivors from British and Irish centers were untraced. Ten years later, however, 23% could not be contacted. Most of them were known to be alive and were traced through the National Health Service Central Register to their current GPs but failed to respond despite 2 written invitations. Unfortunately, research ethics committee approval prohibited any direct contact with the families. Many of these children had not seen their GPs recently, and it was impossible to know whether they had moved and thus never received the letters. There was no difference in loss to follow-up monitoring between the 2 treatment groups, but adverse outcomes might be more prevalent among families that are hardest to trace.21 Future neonatal trials should always obtain parental consent for follow-up monitoring and take steps to facilitate later tracing (such as, in the United Kingdom, recording National Health Service numbers for later retrieval). A sample size of 250 was determined to be required for the original neonatal trial end points of duration of ventilation, oxygen therapy, and hospital stay. However, the numbers traced for follow-up evaluation were reduced by later deaths, lack of follow-up arrangements for some of the centers in the original trial, and failure to contact all eligible families. The final sample of 150 had 80% power to detect a RR of 2.4 for CP or 2.1 for severe disability. It is hoped that, if follow-up data are published from the many other neonatal trials that have not yet reported, then all of these studies can be pooled to yield a reasonable estimate of risk.
Another factor that reduced the power to detect long-term effects of dexamethasone was the overlap in steroid usage between the 2 groups. The original trial allowed for open-label treatment if there was life-threatening deterioration. There was a significant difference in the proportions of children who had received open-label steroids in the neonatal period (39% of the placebo group, compared with 11% of the dexamethasone group). This difference was a measure of the short-term improvement seen with dexamethasone therapy. A secondary analysis performed for the primary outcome measures for the 2 groups, stratified according to centers with no, low, or high open-label steroid usage, showed no clear differences. This type of secondary analysis was performed for the original trial report. Analysis according to treatment received (ie, comparing all children who received trial or open-label steroids with those who did not) would not be informative. Use of open-label steroids in the placebo group was associated with disease severity; children who were more seriously ill were more likely to be given open-label steroids. Therefore, there was an important difference between children who received steroids and those who did not. This would be likely to affect the outcomes, and any differences found might be attributable to differences in disease severity between the groups.22 Overlap between treatment and control groups has been a problem in many neonatal trials.9–11 Indeed, some trials have been designed deliberately to allow open-label treatment after a short course of blinded therapy, which limits seriously their ability to assess later outcomes. It is vital that any future studies apply much stricter control over open-label treatment. The wide use of steroids within trial populations suggests that many participating clinicians were not truly in a state of equipoise but had already decided that the potential benefits of steroids outweighed the potential risks, a view brought into serious question by more recent publications.13,14
Some important questions remain before neonatal corticosteroid treatment is abandoned completely. The apparent anomaly of improved outcomes when steroids are given prenatally6 but increased adverse outcomes with postnatal use among infants of comparable gestational age can be explained most easily by the difference in dosage. Most prenatal studies involve two 12-mg doses (equivalent to a total dose of 0.3–0.4 mg/kg for a mother of 60 to 75 kg, not all of which crosses the placenta23). Neonatal studies have given total doses ranging from 0.89 to 7.8 mg/kg.9–11 Even the lowest doses are many times higher than physiologic concentrations. Concerns about adverse outcomes with neonatal steroid treatment have brought into question the growing practice of administering repeated doses of prenatal steroid treatment to mothers deemed to be at high risk of delivering prematurely. A systematic review of animal studies found that repeated doses have adverse effects on overall fetal growth and on brain growth and function.24 The long-term effects of such a policy in human pregnancies is unclear, and there are several randomized, controlled trials awaiting completion.25
Apart from concerns about total dosage, the question of which steroid is used may be important. Most prenatal studies have involved betamethasone, whereas most neonatal studies have used dexamethasone. Spatial memory in newborn mice exposed to corticosteroid in utero was enhanced after betamethasone treatment but impaired after dexamethasone treatment.26 Dexamethasone has been the preferred corticosteroid for use in reducing cerebral edema because of its ability to cross the blood-brain barrier, which might not have made it the best choice for use in neonatal CLD. A pilot, neonatal, randomized, controlled trial of low-dose hydrocortisone prophylaxis27 showed a reduction in CLD, but a subsequent larger study was abandoned because of an increase in gastrointestinal perforations28 and long-term outcomes are not yet available. Use of inhaled steroids has not been encouraging. A systematic review of 5 randomized trials of corticosteroid administered with metered dose inhalers showed no benefits apart from a reduction in the use of systemically administered steroids,29 and a large trial comparing inhaled budesonide and systemically administered dexamethasone suggested that the inhaled drug was less effective in preventing CLD, although it was safer in terms of other adverse outcomes.30 However, use of nebulized corticosteroid has been associated with earlier extubation31 and may warrant additional study.
The timing of steroid therapy may also be important to the balance of benefits and risks, and this is supported by the difference in findings between the 3 systematic reviews.9–11 The beneficial effects of dexamethasone on lung function are accompanied by a reduction in several markers of pulmonary inflammation,32 the hallmark of neonatal CLD. The group with greatest potential for benefit appears to be patients treated at 7 to 14 days; such treatment avoids the unnecessary risks of giving early prophylactic steroids to infants who might never have developed CLD but optimizes the chance of improvement before the chronic fibrotic changes of bronchopulmonary dysplasia are established. This is the only group of patients for whom systematic review showed a reduced mortality rate,10 with 1 trial recently reporting improved intact survival 15 years after a prolonged course of neonatal steroids.33 Yet it is for this middle group that there have been the fewest trials. An attempt to carry out a well-designed randomized trial with this subgroup, with a smaller dose of dexamethasone and with planned follow-up monitoring (as recommended by the American Academy of Pediatrics and the Canadian Paediatric Society14), was abandoned because of failure to recruit.34 It is unfortunate that the majority of trials, including our own, have been compromised by overenthusiastic use of open-label treatment but a new trial established to test the safety of a lower dosage has failed because of lack of enthusiasm.
Clearly great caution is required in the use of postnatal steroid treatment in CLD, and questions remain regarding whether a lower dosage or a different steroid would be safer than the high doses of dexamethasone used in this and other studies. It is vital for neonatal therapeutic advances that researchers coordinate their efforts to undertake trials with sufficient power and with strict separation between experimental and control groups before treatments with significant unwanted effects enter widespread use.
ACKNOWLEDGMENTS
The original controlled trial, the 3-year follow-up study, and the current study were all funded through the generosity of Action Medical Research.
The Collaborative Dexamethasone Trial Follow-Up Group included: Rosamond Jones, grant-holder and main author; Brenda Strohm, research nurse and trial coordinator; Mary-Grace Breslin and Irene Forster, research nurses; Simon Gates, statistician; Sarah Ayers, computer scientist; Lucy Tully and Ursula Bowler, administrators (National Perinatal Epidemiology Unit); Steering Committee: Peter Brocklehurst (joint grant-holder), Fiona Goddard, and Ann Johnson (National Perinatal Epidemiology Unit); Michael Jones, retired headteacher; Professor Michael Silverman, Leicester Royal Infirmary; Professor Andrew Wilkinson, John Radcliffe Hospital Oxford; Suzanne Dobson and Bonnie Green (Baby Life Support Systems).
Professor Henry Halliday, Royal Victoria Maternity Hospital (Belfast, Ireland), and David Milligan, Royal Victoria Infirmary (Newcastle, United Kingdom), acted as hosts for the research nurses at those centers. We thank the pediatricians who entered patients into the original trial. We thank all of the GPs who forwarded letters to the families and helped in tracing these children and completing questionnaires and also the many teachers who provided information. Above all, our thanks go to all of the parents and teenagers who welcomed the study staff into their homes and gave freely of their time so many years after the original trial.
FOOTNOTES
Accepted Nov 19, 2004.
No conflict of interest declared.
REFERENCES
Horbar JD, McAuliffe TL, Adler SM, et al. Variability in 28-day outcomes for very low birth weight infants: an analysis of 11 neonatal intensive care units. Pediatrics. 1988;82 :554 –559
Sauve RS, Singhal N. Long-term morbidity of infants with bronchopulmonary dysplasia. Pediatrics. 1985;76 :725 –733
Vohr BR, Coll CG, Lobato D, Yunis KA, O'Dea C, Oh W. Neurodevelopmental and medical status of low-birthweight survivors of bronchopulmonary dysplasia at 10 to 12 years of age. Dev Med Child Neurol. 1991;33 :690 –697
Liggins GC, Howie RN. A controlled trial of antepartum glucocorticoid treatment for prevention of the respiratory distress syndrome in premature infants. Pediatrics. 1972;50 :515 –525
Baden M, Bauer CR, Colle E, Klein G, Taeusch HW, Stern L. A controlled trial of hydrocortisone therapy in infants with respiratory distress syndrome. Pediatrics. 1972;50 :526 –534
Crowley P. Prophylactic corticosteroids for preterm birth [Cochrane review]. In: The Cochrane Library. Issue 3. Chichester, United Kingdom: John Wiley & Sons; 2004
Collaborative Dexamethasone Trial Group. Dexamethasone therapy in neonatal chronic lung disease: an international placebo-controlled trial. Pediatrics. 1991;88 :421 –427
Jones R, Wincott E, Elbourne D, Grant A. Controlled trial of dexamethasone in neonatal chronic lung disease: a 3-year follow-up. Pediatrics. 1995;96 :897 –906
Halliday HL, Ehrenkranz RA, Doyle LW. Early postnatal (<96 hours) corticosteroids for preventing chronic lung disease in preterm infants [Cochrane review]. In: The Cochrane Library. Issue 3. Chichester, United Kingdom: John Wiley & Sons; 2004
Halliday HL, Ehrenkranz RA, Doyle LW. Moderately early (7–14 days) postnatal cortico-steroids for preventing chronic lung disease in preterm infants [Cochrane review]. In: The Cochrane Library. Issue 3. Chichester, United Kingdom: John Wiley & Sons; 2004
Halliday HL, Ehrenkranz RA, Doyle LW. Delayed (>3 weeks) postnatal corticosteroids for chronic lung disease in preterm infants [Cochrane review]. In: The Cochrane Library. Issue 3. Chichester, United Kingdom: John Wiley & Sons; 2004
Tarnow-Mordi W, Mitra A. Postnatal dexamethasone in preterm infants. Br Med J. 1999;319 :1385 –1386
American Academy of Pediatrics, Canadian Paediatric Society. Postnatal corticosteroids to treat or prevent chronic lung disease in preterm infants. Pediatrics. 2002;109 :330 –338
Smith P, Hagues N. NFER-Nelson Non-Verbal Reasoning Test Series. Windsor, United Kingdom: NFER-Nelson; 1992
Dunn LM, Dunn LM, Whetton C, Burley J. British Picture Vocabulary Scale. 2nd ed. Windsor, United Kingdom: NFER-Nelson; 1997
Johnson A, Bowler U, Yudkin P, et al. Health and school performance of teenagers born before 29 weeks gestation. Arch Dis Child Fetal Neonatal Ed. 2003;88 :F190 –F198
Goodman R. The Strengths and Difficulties Questionnaire: a research note. J Child Psychol Psychiatry. 1997;38 :581 –586
Yeh TF, Lin YJ, Lin HC, et al. Outcomes at school age after postnatal dexamethasone therapy for lung disease of prematurity. N Engl J Med. 2004;350 :1304 –1313
Tin W, Fritz S, Wariyar U, Hey E. Outcome of very preterm birth: children reviewed with ease at 2 years differ from those followed up with difficulty. Arch Dis Child Fetal Neonatal Ed. 1998;79 :F83 –F87
Rochon J. Issues in adjusting for covariates arising postrandomization in clinical trials. Drug Infect J. 1999;33 :1219 –1228
Ballard PL, Granberg P, Ballard RA. Glucocorticoid levels in maternal and cord serum after prenatal betamethasone therapy to prevent respiratory distress syndrome. J Clin Invest. 1975;56 :1548 –1554
Aghajafari F, Murphy K, Matthews S, Ohlsson A, Amankwah K, Hannah M. Repeated doses of antenatal corticosteroids in animals: a systematic review. Am J Obstet Gynecol. 2002;186 :843 –849
Crowther CA, Harding J. Repeat doses of prenatal corticosteroids for women at risk of preterm birth for preventing neonatal respiratory disease [Cochrane review]. In: The Cochrane Library. Issue 3. Chichester, United Kingdom: John Wiley & Sons; 2004
Rayburn WF, Christensen HD, Gonzalez CL. A placebo-controlled comparison between betamethasone and dexamethasone for fetal maturation: differences in neurobehavioural development of mice offspring. Am J Obstet Gynecol. 1997;176 :842 –851
Watterberg KL, Gerdes JS, Gifford KL, Lin H-M. Prophylaxis against early adrenal insufficiency to prevent chronic lung disease in premature infants. Pediatrics. 1999;104 :1258 –1263
Watterberg KL, Gerdes JS, Cole CH, et al. Prophylaxis of early adrenal insufficiency to prevent bronchopulmonary dysplasia: a multicenter trial. Pediatrics. 2004;114 :1649 –1657
Shah V, Ohlsson A, Halliday HL, Dunn MS. Early administration of inhaled corticosteroids for preventing chronic lung disease in ventilated very low birth weight preterm neonates [Cochrane review]. In: The Cochrane Library. Issue 3. Chichester, United Kingdom: John Wiley & Sons; 2004
Halliday HL, Patterson CC, Halahakoon CWNL. A multicenter, randomized open study of early corticosteroid treatment (OSECT) in preterm infants with respiratory illness: comparison of early and late treatment and of dexamethasone and inhaled budesonide. Pediatrics. 2001;107 :232 –240
Jónsson B, Eriksson M, Sder O, Broberger U, Lagercrantz H. Budesonide delivered by dosimetric jet nebulization to preterm very low birth weight infants at high risk for development of chronic lung disease. Acta Paediatr. 2000;89 :1449 –1455
Yoder MC, Chua R, Tepper R. Effect of dexamethasone on pulmonary inflammation and pulmonary function of ventilator-dependent infants with bronchopulmonary dysplasia. Am Rev Respir Dis. 1991;143 :1044 –1048
Gross SJ, Anbar RD, Mettelman BB. Follow-up at 15 years of preterm infants from a controlled trial of moderately early dexamethasone for the prevention of chronic lung disease. Pediatrics. 2005;115 :681 –687
Doyle L, Davis P, Morley C, McPhee A, Carlin J. Low dose dexamethasone facilitates extubation in ventilator-dependent infants—a multicentre international randomised controlled trial. The DART study investigators . Pediatr Res. 2004;56 :477(Rosamond A. K. Jones, MD,)
ABSTRACT
Objectives. To study neurologic, educational, and psychological status in adolescence of neonates enrolled in a double-blind, randomized, controlled trial of dexamethasone therapy for chronic lung disease.
Participants. A total of 287 infants who were chronically dependent on supplementary oxygen and were 2 to 12 weeks of age were recruited from 31 centers in 6 countries to a randomized, controlled trial of dexamethasone base (0.5 mg/kg per day for 1 week); 95% of survivors were reviewed at 3 years. Survivors from the 25 British and Irish centers were retraced at 13 to 17 years of age.
Outcome Measures. Nonverbal reasoning, British Picture Vocabulary Scale, Goodman Strengths and Difficulties Questionnaire behavior scores, school national test results, teacher ability ratings, and parental and general practitioner questionnaires.
Results. A total of 195 children were eligible for the follow-up study. Information was available for 150 children (77%), with 142 (73%) being assessed in home visits. No baseline differences were detected between the children included in the follow-up study and those not included. There was a slight excess of cerebral palsy in the steroid group, which was not statistically significant (relative risk: 1.58; 95% confidence interval: 0.81–3.07). Overall disability rates in both groups were high (21% moderate and 14% severe), but with no difference between the 2 groups (for severe disability, relative risk: 0.84; 95% confidence interval: 0.37–1.86).
Conclusions. Information was obtained for 150 adolescents randomized to receive dexamethasone or placebo for neonatal chronic lung disease. Rates of disabilities and educational difficulties were high, but with no significant differences between the 2 groups. Some use of open-label steroids in the placebo group plus losses to long-term follow-up monitoring reduced the power of this study to detect clinically important differences, and this study cannot rule out a real increase in cerebral palsy, as reported by others.
Key Words: neonatal chronic lung disease dexamethasone therapy randomized controlled trial developmental follow-up school performance
Abbreviations: CLD, chronic lung disease CP, cerebral palsy GP, general practitioner RR, relative risk CI, confidence interval
Neonatal chronic lung disease (CLD) affects 25% of very low birth weight infants,1 with an increased risk of adverse neurodevelopmental outcomes.2,3 The beneficial effects of prenatal corticosteroid therapy on fetal lung maturation were first shown >30 years ago,4 but a controlled trial of postnatal hydrocortisone therapy for infants with respiratory distress syndrome that was published the same year showed no benefit.5 Additional studies of prenatal corticosteroid therapy have confirmed the reduction in surfactant-deficient respiratory distress syndrome and neonatal mortality rates and secured the place of such therapy in obstetric practice,6 but the place of postnatal steroid treatment remains controversial. A multicenter, randomized, controlled trial of dexamethasone enrolled 287 infants 4 weeks of age with established CLD and showed a significant reduction in the duration of assisted ventilation but no reduction in mortality rates or overall hospital stay.7 Follow-up evaluation of survivors at 3 years showed no persisting benefits but no evidence of increased adverse outcomes,8 and it is these children who form the basis of the current report. A plethora of neonatal trials have resulted in 3 major systematic reviews of corticosteroid treatment, given "early" (age of 0–4 days, 21 trials),9 "moderately early" (age of 7–14 days, 7 trials),10 or "delayed" (age of >3 weeks, 9 trials),11 confirming earlier extubation and a reduction in CLD. Moderately early treatment was also associated with a significant reduction in mortality rates.10 Short-term side effects were common, but most seemed transient.
However, in 1998, a 2-year follow-up study of infants given prophylactic dexamethasone showed an increase in adverse neuromotor outcomes,12 sounding a note of caution. Follow-up data are available for less than one half of the trials included in systematic reviews but have produced conflicting results regarding a possible link with later cerebral palsy (CP). The early group of trials showed an increase in CP rates but no increase in rates of overall disability or combined adverse outcomes of death or CP,9 and the delayed group also showed a nonsignificant increase in CP rates.11 Conversely, moderately early treatment showed a trend for decreased CP rates, but with a smaller number of children being monitored.10 Prenatal steroid use was associated with a significant reduction in neonatal intraventricular hemorrhage rates6 and also a trend toward reduction in CP rates. Individual authors and professional bodies have indicated caution regarding neonatal steroid use and have highlighted the lack of published information on progress into later childhood.13,14 It was these concerns that led to a decision to perform an additional follow-up study in adolescence. The aim was to assess whether early exposure to dexamethasone had any long-term consequences with respect to neurodevelopmental and educational progress, as reported in this article, or growth, lung function, and general health, as reported in the accompanying article.
METHODS
Participants
Between 1984 and 1989, 287 infants with neonatal CLD were recruited to an international, randomized, placebo-controlled trial of dexamethasone therapy.7 Eligible infants were those with no major congenital malformations who were dependent on supplementary oxygen and whose condition was static or deteriorating at 2 to 12 weeks of age (median age: 4 weeks). Randomization was through telephone contact with a central office, with infants stratified according to center and whether they were still receiving assisted ventilation (61%) or not (39%). Treatment was with either dexamethasone phosphate (0.6 mg/kg per day for 1 week, equivalent to 0.5 mg/kg dexamethasone base) or matching saline placebo, given intravenously or orally and with the option of a second tapering 9-day course. Clinicians were allowed to prescribe open-label dexamethasone if there was life-threatening deterioration, but they remained blinded to the trial treatment allocation. The trial was conducted in 31 centers in 6 countries, and a 3-year follow-up study was undertaken.8 Children from continental Europe and North America were excluded from additional follow-up monitoring because of the lack of a reliable system for tracing them. Of children recruited from British and Irish centers, 99% had been successfully evaluated at 3 years of age. No attempt had been made to keep in contact with the families in the intervening years.
Research Ethics Committee Approval
The original, randomized, controlled trial was approved by the research ethics committee at each of the participating centers, and informed parental consent was obtained. Approval for the present study was obtained from the Anglia and Oxford Multicenter Research Ethics Committee, and the appropriate local research ethics committees were informed. The study complied with the ethical guidelines of the British Educational Research Association. Informed consent was obtained from the children and their parents for each aspect of the study. Families have been offered a summary of the overall findings of the study.
Tracing and Contact
Initial contact was made with the last known general practitioners (GPs), to check whether the children were still registered with them. A search for the remainder of the children was made through the National Health Service Central Register, to ascertain any deaths, and the Family Health Services Authority, to determine the current GPs for survivors. A letter was sent requesting the GPs to forward details of the study to the families. A reply slip and prepaid envelope to the study office were enclosed. If parents failed to respond, then a second mailing was sent. For continued nonresponders, GPs were approached and asked whether there was any known reason for nonresponse; they were asked to flag the patients' notes and to speak to them about the study if they attended the surgery. No direct contact was made with families without their agreement.
Consideration was given to obtaining consent from the children themselves, who were then adolescents. A pilot study of 67 families ascertained that 58 children were already aware of their preterm birth and of the trial and no parent anticipated difficulties in talking to the child about it. In addition to the explanatory letter to parents, a separate information leaflet was enclosed for parents to give to the child. This leaflet was drafted with input from a group of healthy teenagers and also parents of formerly preterm infants. Consent was obtained from parents and children for each part of the study separately. Therefore, families could agree to a home visit but decline contact with their school. Family practitioners, with the families' consent, were given the results of nurses' assessments but not the teachers' reports. All home visits were conducted between September 2001 and December 2002.
Assessment Tools
All questionnaires and information leaflets are available from the author or electronically (www.npeu.ox.ac.uk/dex). The families completed a questionnaire on functional status, diagnoses of potentially disabling conditions (visual or hearing impairments, learning disabilities, CP, and epilepsy), and the child's schooling. Questions on respiratory symptoms are detailed in the accompanying article. GPs also completed a questionnaire, reporting on any known functional problems, diagnoses, and hospital admissions. No specific definitions were provided, and the diagnosis of CP was made by the pediatrician responsible for each child's care. Children were visited at home by 1 of 3 research nurses, who were blinded to the children's original treatment allocation. To assess different areas of cognitive ability, a nonverbal reasoning test15 and the British Picture Vocabulary Scale16 were administered. These tests, published by the National Foundation for Educational Research, are used widely for British schoolchildren and have both been restandardized recently. Training was provided by staff members from the National Foundation for Educational Research. Results are expressed as a performance age, which is compared with the child's actual age to yield a quotient. The tests use a multiple-choice type of approach, and the nonverbal reasoning test can be completed by a child with poor reading skills or one with poor coordination or CP, because no drawing or fine motor skills are involved. The tests do not need to be administered by a trained psychologist. These tests could not be administered to 2 blind children, and information regarding their cognitive abilities was obtained from their specialist teachers.
Disability Rating
For the purposes of an overall disability rating, the results of the nonverbal reasoning test and the British Picture Vocabulary Scale were averaged as a proxy for IQ. Moderate disability was defined as 1 or 2 of the following: IQ 2 to 3 SDs below the mean, ambulatory CP, hearing deficits corrected with aids, impaired vision, or behavior disorder with a major impact on schooling. Severe disability was any of the following: IQ >3 SDs below the mean, wheelchair-dependent CP, uncorrectable hearing loss, blind (perception of light only), or 3 moderate disabilities.
School Assessment
Permission was sought from parents and children to contact their schools for completion of a teacher assessment form, including the child's most recent national tests results (Key Stage 2 at age 11 or Key Stage 3 at age 14 for England and Wales or the equivalent examinations for Scotland and Ireland). Teachers were also asked to estimate the child's position within an average mixed-ability class for a range of subjects, on an analog scale of 1 to 30 (with 1 being the lowest). The same questionnaire was used successfully for another cohort of preterm infants evaluated as teenagers.17 The teacher assessment also incorporated a standardized behavior inventory, the Strengths and Difficulties Questionnaire described by Goodman.18 The questionnaire includes an "impact score" summarizing the effects of the child's difficulties on the individual child, peer relationships, the teacher, and the whole class. Teachers were asked to leave no details of the study in the child's school records when they returned the questionnaire.
Sample Size, Outcome Measures, and Statistical Methods
The sample size was predetermined by the size required for the original randomized trial minus any deaths and enrollments from centers outside Britain and Ireland. Primary outcome measures were the proportions of children >2 SD below the mean for nonverbal reasoning or British Picture Vocabulary Scale scores. A sample size of 195 yielded an 80% chance of showing an increase in incidence in the primary outcomes from 20% to 39% or a decrease to 6%. Secondary outcome measures included the proportions of children with CP, sensory impairments, or severe neurosensory disabilities, school national test results, teacher ratings, and results of the Strengths and Difficulties Questionnaire. Statistical analyses were conducted with SPSS version 11 software (SPSS, Chicago, IL), with results given as relative risks (RRs) or differences with 95% confidence intervals (CIs); the 2 test for trend was used when appropriate. All analyses were performed on an intention-to-treat basis. Centers were ranked according to their use of open-label steroids, and a subanalysis was performed on the main outcome measures comparing children from the 11 centers where no open-label treatment was used with children from centers with low use (1–65%) and those from centers with high use (66–100%).
RESULTS
Of the original 287 infants randomized, 195 were eligible for follow-up assessments (Fig 1). Five children were known to have spastic quadriplegia and severe learning difficulties, and their GPs or health visitors provided information at 3 years. None have died since then, but they were not contacted for this later follow-up evaluation because it was thought that they would be unable to complete any of the test items. These children are included in the results on disabilities. Of the 190 children eligible for contact, 3 children were known to have emigrated, 1 child was known to have gone into foster care at the 3-year follow-up evaluation, and 1 other child who was evaluated at 3 years could not be traced at all. Thirty-four families failed to respond despite reminder letters, and 6 declined to participate. Health questionnaires were completed by 145 (76%) of the 190 families, a home visit was completed for 142 (75%), and a teacher questionnaire was returned for 118 children (63%). Including the 5 severely disabled children, some information was available for 150 (77%) of 195 survivors. There was no evidence of any differences between the children included in the follow-up study and those not included (Table 1). There was also no evidence that the comparison of the dexamethasone and placebo groups differed between those included in the follow-up study and those not included (Table 1). In the neonatal trial, placebo-treated infants were more likely to receive open-label steroids than were infants in the dexamethasone group (11% vs 39%), which reflects the short-term benefits of steroids. However, there was no difference in open-label steroid usage between those evaluated in adolescence and those lost to follow-up monitoring.
The key neurosensory outcomes are shown in Table 2. Nonverbal reasoning scores and vocabulary scores showed no significant differences between the 2 groups either in the proportions of children >2 SDs below the mean or in the mean values themselves. Twenty-nine children had a diagnosis of CP (dexamethasone group: 17 of 71 subjects; placebo: 12 of 79 subjects; RR: 1.58; 95% CI: 0.81–3.07). There was no difference between the groups in overall disability, whether moderate or severe, in the use of therapists, or in the incidence of seizures or hydrocephalus.
Teachers completed the Goodman Strengths and Difficulties Questionnaire (Table 4). In an average population of children, 10% are expected to have abnormal scores and an additional 10% to have borderline scores.18 These children showed an increased proportion with abnormal scores but no difference between the 2 groups, in either total scores (22% abnormal, as shown in Table 4) or individual subscores (conduct problems, emotional symptoms, hyperactivity, and peer problems; data not shown). Teachers reported some difficulties in behavior, emotion, or concentration for more than one half of the children, with these difficulties significantly affecting the child and the rest of the class. However, teachers reported better than average scores for prosocial behavior.
Use of open-label steroids for 39% of the placebo-treated children and 11% of the dexamethasone-treated children had the potential to reduce any differences between the 2 groups. A secondary analysis was therefore performed for the primary outcome measures for the 2 groups, stratified according to centers with no, low, or high open-label steroid usage (Table 5). For children from centers using no open-label steroids, the RRs of low nonverbal reasoning test scores, low British Picture Vocabulary Scale scores, and CP were greater than those for children from centers with low or high steroid usage, but all CIs overlapped 0, as did the ratio of RRs in interaction tests, which suggests that all of these differences could have occurred by chance.
DISCUSSION
This is the largest follow-up study through to adolescence of preterm infants treated with neonatal corticosteroid therapy. The results confirmed the high risk of adverse outcomes among infants with neonatal CLD; 35% had moderate or severe disabilities, including 19% with a diagnosis of CP. Twenty-one percent attended a "special needs" facility and two thirds were achieving below expected levels for their age in national educational assessments. Despite their difficulties, most children were seen as cooperative and considerate members of their schools, with high ratings on the prosocial behavior scale. However, no clear differences were seen between the 2 groups for any outcome measure. This is in contrast to reports for younger children given early neonatal dexamethasone treatment, which showed a significantly higher incidence of CP9 and a finding of lower IQ at 8 years of age.20 There are various possible explanations for these differences.
The first concern involves tracing of the children. At the 3-year follow-up assessment, only 1% of survivors from British and Irish centers were untraced. Ten years later, however, 23% could not be contacted. Most of them were known to be alive and were traced through the National Health Service Central Register to their current GPs but failed to respond despite 2 written invitations. Unfortunately, research ethics committee approval prohibited any direct contact with the families. Many of these children had not seen their GPs recently, and it was impossible to know whether they had moved and thus never received the letters. There was no difference in loss to follow-up monitoring between the 2 treatment groups, but adverse outcomes might be more prevalent among families that are hardest to trace.21 Future neonatal trials should always obtain parental consent for follow-up monitoring and take steps to facilitate later tracing (such as, in the United Kingdom, recording National Health Service numbers for later retrieval). A sample size of 250 was determined to be required for the original neonatal trial end points of duration of ventilation, oxygen therapy, and hospital stay. However, the numbers traced for follow-up evaluation were reduced by later deaths, lack of follow-up arrangements for some of the centers in the original trial, and failure to contact all eligible families. The final sample of 150 had 80% power to detect a RR of 2.4 for CP or 2.1 for severe disability. It is hoped that, if follow-up data are published from the many other neonatal trials that have not yet reported, then all of these studies can be pooled to yield a reasonable estimate of risk.
Another factor that reduced the power to detect long-term effects of dexamethasone was the overlap in steroid usage between the 2 groups. The original trial allowed for open-label treatment if there was life-threatening deterioration. There was a significant difference in the proportions of children who had received open-label steroids in the neonatal period (39% of the placebo group, compared with 11% of the dexamethasone group). This difference was a measure of the short-term improvement seen with dexamethasone therapy. A secondary analysis performed for the primary outcome measures for the 2 groups, stratified according to centers with no, low, or high open-label steroid usage, showed no clear differences. This type of secondary analysis was performed for the original trial report. Analysis according to treatment received (ie, comparing all children who received trial or open-label steroids with those who did not) would not be informative. Use of open-label steroids in the placebo group was associated with disease severity; children who were more seriously ill were more likely to be given open-label steroids. Therefore, there was an important difference between children who received steroids and those who did not. This would be likely to affect the outcomes, and any differences found might be attributable to differences in disease severity between the groups.22 Overlap between treatment and control groups has been a problem in many neonatal trials.9–11 Indeed, some trials have been designed deliberately to allow open-label treatment after a short course of blinded therapy, which limits seriously their ability to assess later outcomes. It is vital that any future studies apply much stricter control over open-label treatment. The wide use of steroids within trial populations suggests that many participating clinicians were not truly in a state of equipoise but had already decided that the potential benefits of steroids outweighed the potential risks, a view brought into serious question by more recent publications.13,14
Some important questions remain before neonatal corticosteroid treatment is abandoned completely. The apparent anomaly of improved outcomes when steroids are given prenatally6 but increased adverse outcomes with postnatal use among infants of comparable gestational age can be explained most easily by the difference in dosage. Most prenatal studies involve two 12-mg doses (equivalent to a total dose of 0.3–0.4 mg/kg for a mother of 60 to 75 kg, not all of which crosses the placenta23). Neonatal studies have given total doses ranging from 0.89 to 7.8 mg/kg.9–11 Even the lowest doses are many times higher than physiologic concentrations. Concerns about adverse outcomes with neonatal steroid treatment have brought into question the growing practice of administering repeated doses of prenatal steroid treatment to mothers deemed to be at high risk of delivering prematurely. A systematic review of animal studies found that repeated doses have adverse effects on overall fetal growth and on brain growth and function.24 The long-term effects of such a policy in human pregnancies is unclear, and there are several randomized, controlled trials awaiting completion.25
Apart from concerns about total dosage, the question of which steroid is used may be important. Most prenatal studies have involved betamethasone, whereas most neonatal studies have used dexamethasone. Spatial memory in newborn mice exposed to corticosteroid in utero was enhanced after betamethasone treatment but impaired after dexamethasone treatment.26 Dexamethasone has been the preferred corticosteroid for use in reducing cerebral edema because of its ability to cross the blood-brain barrier, which might not have made it the best choice for use in neonatal CLD. A pilot, neonatal, randomized, controlled trial of low-dose hydrocortisone prophylaxis27 showed a reduction in CLD, but a subsequent larger study was abandoned because of an increase in gastrointestinal perforations28 and long-term outcomes are not yet available. Use of inhaled steroids has not been encouraging. A systematic review of 5 randomized trials of corticosteroid administered with metered dose inhalers showed no benefits apart from a reduction in the use of systemically administered steroids,29 and a large trial comparing inhaled budesonide and systemically administered dexamethasone suggested that the inhaled drug was less effective in preventing CLD, although it was safer in terms of other adverse outcomes.30 However, use of nebulized corticosteroid has been associated with earlier extubation31 and may warrant additional study.
The timing of steroid therapy may also be important to the balance of benefits and risks, and this is supported by the difference in findings between the 3 systematic reviews.9–11 The beneficial effects of dexamethasone on lung function are accompanied by a reduction in several markers of pulmonary inflammation,32 the hallmark of neonatal CLD. The group with greatest potential for benefit appears to be patients treated at 7 to 14 days; such treatment avoids the unnecessary risks of giving early prophylactic steroids to infants who might never have developed CLD but optimizes the chance of improvement before the chronic fibrotic changes of bronchopulmonary dysplasia are established. This is the only group of patients for whom systematic review showed a reduced mortality rate,10 with 1 trial recently reporting improved intact survival 15 years after a prolonged course of neonatal steroids.33 Yet it is for this middle group that there have been the fewest trials. An attempt to carry out a well-designed randomized trial with this subgroup, with a smaller dose of dexamethasone and with planned follow-up monitoring (as recommended by the American Academy of Pediatrics and the Canadian Paediatric Society14), was abandoned because of failure to recruit.34 It is unfortunate that the majority of trials, including our own, have been compromised by overenthusiastic use of open-label treatment but a new trial established to test the safety of a lower dosage has failed because of lack of enthusiasm.
Clearly great caution is required in the use of postnatal steroid treatment in CLD, and questions remain regarding whether a lower dosage or a different steroid would be safer than the high doses of dexamethasone used in this and other studies. It is vital for neonatal therapeutic advances that researchers coordinate their efforts to undertake trials with sufficient power and with strict separation between experimental and control groups before treatments with significant unwanted effects enter widespread use.
ACKNOWLEDGMENTS
The original controlled trial, the 3-year follow-up study, and the current study were all funded through the generosity of Action Medical Research.
The Collaborative Dexamethasone Trial Follow-Up Group included: Rosamond Jones, grant-holder and main author; Brenda Strohm, research nurse and trial coordinator; Mary-Grace Breslin and Irene Forster, research nurses; Simon Gates, statistician; Sarah Ayers, computer scientist; Lucy Tully and Ursula Bowler, administrators (National Perinatal Epidemiology Unit); Steering Committee: Peter Brocklehurst (joint grant-holder), Fiona Goddard, and Ann Johnson (National Perinatal Epidemiology Unit); Michael Jones, retired headteacher; Professor Michael Silverman, Leicester Royal Infirmary; Professor Andrew Wilkinson, John Radcliffe Hospital Oxford; Suzanne Dobson and Bonnie Green (Baby Life Support Systems).
Professor Henry Halliday, Royal Victoria Maternity Hospital (Belfast, Ireland), and David Milligan, Royal Victoria Infirmary (Newcastle, United Kingdom), acted as hosts for the research nurses at those centers. We thank the pediatricians who entered patients into the original trial. We thank all of the GPs who forwarded letters to the families and helped in tracing these children and completing questionnaires and also the many teachers who provided information. Above all, our thanks go to all of the parents and teenagers who welcomed the study staff into their homes and gave freely of their time so many years after the original trial.
FOOTNOTES
Accepted Nov 19, 2004.
No conflict of interest declared.
REFERENCES
Horbar JD, McAuliffe TL, Adler SM, et al. Variability in 28-day outcomes for very low birth weight infants: an analysis of 11 neonatal intensive care units. Pediatrics. 1988;82 :554 –559
Sauve RS, Singhal N. Long-term morbidity of infants with bronchopulmonary dysplasia. Pediatrics. 1985;76 :725 –733
Vohr BR, Coll CG, Lobato D, Yunis KA, O'Dea C, Oh W. Neurodevelopmental and medical status of low-birthweight survivors of bronchopulmonary dysplasia at 10 to 12 years of age. Dev Med Child Neurol. 1991;33 :690 –697
Liggins GC, Howie RN. A controlled trial of antepartum glucocorticoid treatment for prevention of the respiratory distress syndrome in premature infants. Pediatrics. 1972;50 :515 –525
Baden M, Bauer CR, Colle E, Klein G, Taeusch HW, Stern L. A controlled trial of hydrocortisone therapy in infants with respiratory distress syndrome. Pediatrics. 1972;50 :526 –534
Crowley P. Prophylactic corticosteroids for preterm birth [Cochrane review]. In: The Cochrane Library. Issue 3. Chichester, United Kingdom: John Wiley & Sons; 2004
Collaborative Dexamethasone Trial Group. Dexamethasone therapy in neonatal chronic lung disease: an international placebo-controlled trial. Pediatrics. 1991;88 :421 –427
Jones R, Wincott E, Elbourne D, Grant A. Controlled trial of dexamethasone in neonatal chronic lung disease: a 3-year follow-up. Pediatrics. 1995;96 :897 –906
Halliday HL, Ehrenkranz RA, Doyle LW. Early postnatal (<96 hours) corticosteroids for preventing chronic lung disease in preterm infants [Cochrane review]. In: The Cochrane Library. Issue 3. Chichester, United Kingdom: John Wiley & Sons; 2004
Halliday HL, Ehrenkranz RA, Doyle LW. Moderately early (7–14 days) postnatal cortico-steroids for preventing chronic lung disease in preterm infants [Cochrane review]. In: The Cochrane Library. Issue 3. Chichester, United Kingdom: John Wiley & Sons; 2004
Halliday HL, Ehrenkranz RA, Doyle LW. Delayed (>3 weeks) postnatal corticosteroids for chronic lung disease in preterm infants [Cochrane review]. In: The Cochrane Library. Issue 3. Chichester, United Kingdom: John Wiley & Sons; 2004
Tarnow-Mordi W, Mitra A. Postnatal dexamethasone in preterm infants. Br Med J. 1999;319 :1385 –1386
American Academy of Pediatrics, Canadian Paediatric Society. Postnatal corticosteroids to treat or prevent chronic lung disease in preterm infants. Pediatrics. 2002;109 :330 –338
Smith P, Hagues N. NFER-Nelson Non-Verbal Reasoning Test Series. Windsor, United Kingdom: NFER-Nelson; 1992
Dunn LM, Dunn LM, Whetton C, Burley J. British Picture Vocabulary Scale. 2nd ed. Windsor, United Kingdom: NFER-Nelson; 1997
Johnson A, Bowler U, Yudkin P, et al. Health and school performance of teenagers born before 29 weeks gestation. Arch Dis Child Fetal Neonatal Ed. 2003;88 :F190 –F198
Goodman R. The Strengths and Difficulties Questionnaire: a research note. J Child Psychol Psychiatry. 1997;38 :581 –586
Yeh TF, Lin YJ, Lin HC, et al. Outcomes at school age after postnatal dexamethasone therapy for lung disease of prematurity. N Engl J Med. 2004;350 :1304 –1313
Tin W, Fritz S, Wariyar U, Hey E. Outcome of very preterm birth: children reviewed with ease at 2 years differ from those followed up with difficulty. Arch Dis Child Fetal Neonatal Ed. 1998;79 :F83 –F87
Rochon J. Issues in adjusting for covariates arising postrandomization in clinical trials. Drug Infect J. 1999;33 :1219 –1228
Ballard PL, Granberg P, Ballard RA. Glucocorticoid levels in maternal and cord serum after prenatal betamethasone therapy to prevent respiratory distress syndrome. J Clin Invest. 1975;56 :1548 –1554
Aghajafari F, Murphy K, Matthews S, Ohlsson A, Amankwah K, Hannah M. Repeated doses of antenatal corticosteroids in animals: a systematic review. Am J Obstet Gynecol. 2002;186 :843 –849
Crowther CA, Harding J. Repeat doses of prenatal corticosteroids for women at risk of preterm birth for preventing neonatal respiratory disease [Cochrane review]. In: The Cochrane Library. Issue 3. Chichester, United Kingdom: John Wiley & Sons; 2004
Rayburn WF, Christensen HD, Gonzalez CL. A placebo-controlled comparison between betamethasone and dexamethasone for fetal maturation: differences in neurobehavioural development of mice offspring. Am J Obstet Gynecol. 1997;176 :842 –851
Watterberg KL, Gerdes JS, Gifford KL, Lin H-M. Prophylaxis against early adrenal insufficiency to prevent chronic lung disease in premature infants. Pediatrics. 1999;104 :1258 –1263
Watterberg KL, Gerdes JS, Cole CH, et al. Prophylaxis of early adrenal insufficiency to prevent bronchopulmonary dysplasia: a multicenter trial. Pediatrics. 2004;114 :1649 –1657
Shah V, Ohlsson A, Halliday HL, Dunn MS. Early administration of inhaled corticosteroids for preventing chronic lung disease in ventilated very low birth weight preterm neonates [Cochrane review]. In: The Cochrane Library. Issue 3. Chichester, United Kingdom: John Wiley & Sons; 2004
Halliday HL, Patterson CC, Halahakoon CWNL. A multicenter, randomized open study of early corticosteroid treatment (OSECT) in preterm infants with respiratory illness: comparison of early and late treatment and of dexamethasone and inhaled budesonide. Pediatrics. 2001;107 :232 –240
Jónsson B, Eriksson M, Sder O, Broberger U, Lagercrantz H. Budesonide delivered by dosimetric jet nebulization to preterm very low birth weight infants at high risk for development of chronic lung disease. Acta Paediatr. 2000;89 :1449 –1455
Yoder MC, Chua R, Tepper R. Effect of dexamethasone on pulmonary inflammation and pulmonary function of ventilator-dependent infants with bronchopulmonary dysplasia. Am Rev Respir Dis. 1991;143 :1044 –1048
Gross SJ, Anbar RD, Mettelman BB. Follow-up at 15 years of preterm infants from a controlled trial of moderately early dexamethasone for the prevention of chronic lung disease. Pediatrics. 2005;115 :681 –687
Doyle L, Davis P, Morley C, McPhee A, Carlin J. Low dose dexamethasone facilitates extubation in ventilator-dependent infants—a multicentre international randomised controlled trial. The DART study investigators . Pediatr Res. 2004;56 :477(Rosamond A. K. Jones, MD,)