Randomized Placebo-controlled Trial of a 42-Day Tapering Course of Dexamethasone to Reduce the Duration of Ventilator Dependency in Very Low Birth Weight Infants: Outcome of Study Participants at 1-Year Adjusted Age
Objective. Ventilator-dependent preterm infants are often treated with a prolonged tapering course of dexamethasone to decrease the risk and severity of chronic lung disease. The objective of this study was to assess the effect of this therapy on developmental outcome at 1 year of age.
Methods. Study participants were 118 very low birth weight infants who, at 15 to 25 days of life, were not weaning from assisted ventilation and were then enrolled in a randomized, placebo-controlled, double-blind trial of a 42-day tapering course of dexamethasone. Infants were examined at 1 year of age, adjusted for prematurity, by a pediatrician and a child psychologist. A physical and neurologic examination was performed, and the Bayley Scales of Infant Development were administered. All examiners were blind to treatment group.
Results. Groups were similar in terms of birth weight, gestational age, gender, and race. A higher percentage of dexamethasone recipients had major intracranial abnormalities diagnosed by ultrasonography (21% vs 11%). Group differences were not found for Bayley Mental Development Index (median [range] for dexamethasone-treated group, 94 [50–123]; for placebo group, 90 [28–117]) or Psychomotor Development Index Index (median [range]) for dexamethasone-treated group, 78 (50–109); for placebo-treated group, 81 [28–117]). More dexamethasone-treated infants had cerebral palsy (25% vs 7%) and abnormal neurologic examination findings (45% vs 16%). In stratified analyses, adjusted for major cranial ultrasound abnormalities, these associations persisted (OR values for cerebral palsy, 5.3; 95% CI: 1.3–21.4; OR values for neurologic abnormality 3.6; 95% CI: 1.2–11.0).
Conclusions. A 42-day tapering course of dexamethasone was associated with an increased risk of cerebral palsy. Possible explanations include an adverse effect of this therapy on brain development and/or improved survival of infants who either already have neurologic injury or who are at increased risk for such injury.
- bronchopulmonary dysplasia
- chronic lung disease
- very low birth weight
- randomized controlled trials
- developmental outcome
- cerebral palsy
Randomized, controlled trials have indicated that in ventilator-dependent preterm infants, dexamethasone improves pulmonary function and decreases the duration of ventilator dependency.1–11 The effect of this drug on long-term growth and development has not been well-studied. In a randomized trial by Cummings and associates,1 infants treated with a 42-day course of dexamethasone were more likely to have normal outcome at 15 months of age, compared with control infants and infants treated with an 18-day course of dexamethasone. This study, however, involved a small number of infants and an important confounding factor—severe intracranial hemorrhage—occurred in 3 of 5 control infants but in only 2 of 9 infants treated with a 42-day course of dexamethasone. In two studies involving larger samples and historical controls, no difference in outcome was found between dexamethasone-treated infants and control infants.12 ,13 Follow-up of the participants in the largest reported trial of dexamethasone indicated no effect of dexamethasone on outcome at 3 years of age,14 but the statistical power of this study was limited because 40% of infants randomized to placebo were later treated with open-label dexamethasone. More recently, Yeh and colleagues reported that dexamethasone-treated infants had a higher incidence of neuromotor dysfunction at age 2 years.15
Here we describe the outcome at 1-year adjusted age for 95 surviving infants who participated in a randomized, placebo-controlled trial of a 42-day tapering course of dexamethasone. In this trial, crossing over (dexamethasone treatment of infants randomized to placebo) was not permitted.
Randomized Placebo-controlled Trial of Dexamethasone
The study protocol was approved by the institutional review board of Wake Forest University School of Medicine. Details of the randomized controlled trial of dexamethasone have been described in an accompanying article. Briefly, study subjects were born between April 1992 and May 1995, and received intensive care in the two level III neonatal intensive care units where clinical care is supervised by faculty neonatologists of Wake Forest University School of Medicine. Infants were eligible for randomization if they met the following criteria: 1) birth weight < 1501 g; 2) age between 15 and 25 days; 3) <10% decrease in ventilator settings for previous 24 hours and Fio2 ≥ 0.3; 4) no clinical signs of sepsis; and 5) an echocardiogram indicating the absence of a patent ductus arteriosus. Infants randomized to dexamethasone were treated according to the following schedule: 0.25 mg/kg bid for 3 days, then 0.15 mg/kg bid for 3 days, then a 10% reduction in the dose every 3 days until the dose of 0.1 mg/kg was reached on day 34. After 3 days on this dose, 0.1 mg/kg qod was given until 42 days after entry. Infants randomized to placebo were treated with an equivalent volume of vehicle.
Neonatal cranial ultrasound examinations were performed as a routine aspect of clinical care at 4 to 7 days and at 14 days of age. Additional ultrasound studies were performed at the discretion of the attending neonatologist. Retinal examinations were performed at 4 to 6 weeks of life by a single pediatric ophthalmologist (R.G.W.) and were repeated at 1 to 2 week intervals until healing or retinal detachment was noted. Laser surgery for retinopathy of prematurity was performed when grade 3 retinopathy with plus disease was diagnosed with at least 5 contiguous clock hours or 8 noncontiguous clock hours of retinal involvement.16
Infants who survived to discharge from the hospital were enrolled in an infant follow-up project that began in 1976 and has operated continuously since, with funding from the North Carolina Department of Environment, Health, and Natural Resources. The methods used by this project to track and evaluate participants have been described in detail previously.17 Briefly, families are contacted by letter and/or telephone in the 1st month after their infant's discharge, and when the infant is 6 and 9 months' adjusted age. At 12 months' adjusted age, infants are scheduled for a multidisciplinary evaluation at the Wake Forest University School of Medicine Developmental Evaluation Clinic. This evaluation consists of a hearing screen, administration of the Bayley Scales of Infant Development18 and the Vineland Adaptive Behavioral Scales19 by clinical child psychologists, and a physical and neurologic examination by a developmental pediatrician (K.L.K.) or a neonatologist with special interest in developmental follow-up (T.M.O. and R.G.D.). Infants tested through October 1993 (n = 10) were administered the original Bayley Scales of Infant Development18; those tested after October 1993 were administered the Bayley Scales of Infant Development–Second Edition.20 Examining pediatricians were aware of each infant's medical history at the time of examination. Psychologists were not aware of the medical history until completion of developmental testing, at which time they learned the infant's gestational age at birth. Infants suspected of having a neurologic abnormality were also seen by a pediatric physical therapist. Cerebral palsy was diagnosed only if both a pediatrician and a physical therapist agreed on the presence of abnormal control of movement and posture, with impaired motor function. Based on subjective assessment of the examining physician, the neurologic examination was classified as indicative of no abnormality, mild abnormality, or moderate to severe abnormality. Infants who were felt by the pediatrician and physical therapist to have hypotonia (which typically was most prominent in the truncal musculature) were classified as having mild neurologic abnormality but were not classified as having cerebral palsy. Thus, the group of infants classified as having a neurologic abnormality included infants with possible or definite cerebral palsy as well as infants with hypotonia.
Chronic lung disease was defined as the use of supplemental oxygen at 36 weeks' postconception.21 The definitions described by Hyde22 and Heneghan23 were used to categorize infants according to the predominant radiographic finding indicative of chronic lung disease. Small for gestational age was defined as a birth weight less than the 10th percentile, based on data reported by Lubchenco.24 The assignment of gestational age was based on the date of the mother's last menstrual period unless this was not available, in which case an obstetrician's estimate was used. When no prenatal estimate was available, gestational age was based on assessment of the neonate.25 Based on work by Stewart and colleagues,26 we defined a major cranial ultrasound abnormality as 1) subependymal/intraventricular hemorrhage with posthemorrhagic hydrocephalus requiring placement of a shunt; 2) persistent but nonprogressive ventricular dilatation; or 3) intraparenchymal echodensity or echolucency in the periventricular white matter (consistent with periventricular hemorrhagic infarction or periventricular leukomalacia).27 The reliability of interpretations of cranial ultrasound examinations at our medical center has been reported previously.28 In general agreement with others,29 ,30 we defined a major neurosensory impairment as a Mental Development Index >2 SD units below the mean (68 for the original Bayley Scales of Infant Development18 and 70 for the Bayley Scales of Infant Development–Second Edition),20) cerebral palsy, or blindness, as diagnosed by a pediatric ophthalmologist.
Associations between dexamethasone treatment and dichotomous outcomes were described as ORs with exact 95% CI. Group comparisons were performed using the Wilcoxon rank sum test for continuous variables and Fisher's exact test for categorical variables. The results of stratified analyses were expressed as Mantel-Haenszel summary odds ratios.31 For multivariate analyses involving categorical outcomes, logistic regression was used. For the outcome of cerebral palsy, which was categorized as absent, possibly present, or definitely present, multivariate analyses were performed in three ways. The conclusions from multivariate analyses were the same regardless of whether cerebral palsy was modeled as a trichotomous outcome (using polychotomous logistic regression), as a dichotomy, grouping “possible” cases with “definite” cases, or as a dichotomy with “possible” cases excluded from the analysis. Analyses were performed using the microcomputer programs PC-SAS32 and StatXact.33
Characteristics of Study Infants
Fifty-seven infants were randomized to receive a 42-day tapering course of dexamethasone; 61 to receive placebo. Fifty (88%) dexamethasone-treated infants and 45 (74%) placebo recipients survived to 1 year adjusted age (OR for mortality, dexamethasone:placebo = 0.4 [95% CI: 0.1–1.0]; P = .066). Attributes of surviving infants are shown in Table 1. All but 1 infant received surfactant. Birth weight, gestational age, gender, race, maternal education, maternal age, and marital status were similar for the 2 groups. Antenatal steroids had been given to 48% of mothers of dexamethasone-treated infants, compared with 16% of mothers of placebo recipients (P = .001).
As shown in Fig 1, major cranial ultrasound abnormalities, defined as posthemorrhagic hydrocephalus, persistent ventricular dilatation, or intraparenchymal (intracerebral) echodensity or echolucency, were present on cranial ultrasound in 10 (20%) dexamethasone-treated infants, compared with 5 (11%) placebo recipients. Before randomization, 5 dexa-methasone-treated infants (10%) had major cranial ultrasound abnormalities, and an additional 5 infants developed such abnormalities after randomization and treatment with dexamethasone. Among placebo recipients, 4 (8%) had major abnormalities before randomization. After randomization, intraparenchymal echolucency was found in 1 other infant who had not undergone ultrasonography before randomization.
The parents of 2 infants treated with dexamethasone refused to bring their infant for an evaluation at 1 year of age, even after they were offered a financial incentive. According to parent report, these 2 infants (2% of the study cohort) were alive at 1 year of age. Two other infants were evaluated at state-supported (North Carolina) Developmental Evaluation Clinics, and information was provided to us by these centers. One infant, who was treated with dexamethasone, was diagnosed as having possible cerebral palsy and definite developmental delay. The other, a placebo recipient, was classified as free from cerebral palsy but was thought to have borderline developmental delay.
Outcome at 1 Year Adjusted Age
The information provided in Tables 2and 3 is based on the 93 infants who were evaluated either at our medical center (47 dexamethasone recipients and 44 placebo recipients) and at other state- supported Developmental Evaluation Clinics (1 placebo recipient and 1 dexamethasone recipient). For analyses involving either cerebral palsy or neurologic abnormality (cerebral palsy or hypotonia) as the outcome of interest, conclusions were the same, regardless of whether the 2 infants evaluated outside of our medical center were included or excluded.
Dexamethasone-treated infants and placebo recipients had similar scores on the Bayley Scales of Infant Development for both Mental Development Index and Psychomotor Development Index, and on the Vineland Adaptive Behavioral Scales (Table 2). A higher proportion of dexamethasone recipients were diagnosed with cerebral palsy (12/48 [25%] vs 3/45 [7%]; OR, 4.6; 95% CI: 1.3–21.7; P = .02), and a higher proportion were classified as having a neurologic abnormality (20/48 [42%] vs 8/45 [18%]; OR, 3.3; 95% CI: 1.3–9.0;P = .01). If cerebral palsy and neurologic abnormality were modeled as a trichotomous outcomes (no cerebral palsy, suspect cerebral palsy, definite cerebral palsy, and no neurologic abnormality, mild abnormality, and moderate to severe abnormality), the difference between dexamethasone-treated and placebo groups remained significant (cerebral palsy, P = .006; neurologic abnormality,P = .03; χ2 tests for trend).
When infants were stratified according to the presence or absence of major cranial ultrasound abnormality, as defined previously, the association between dexamethasone receipt and a higher risk of cerebral palsy and neurologic abnormality persisted (Table 4). Excluding infants with possible cerebral palsy, ORs for the association of dexamethasone and cerebral palsy, adjusting for major cranial ultrasound abnormality, was 16.0 (95% CI: 1.3–200.8; P = .01). Combining infants with possible cerebral palsy with those with definite cerebral palsy yielded an OR of 5.3 (95% CI: 1.3–21.4; P = .01).
To assess possible confounding attributable to factors other than major cranial ultrasound abnormality, we examined univariate associations between each of the factors listed in Table 1 and the presence of cerebral palsy. Only three factors were associated with cerebral palsy at the 0.1 level of significance: birth weight, dexamethasone treatment, and major cranial ultrasound abnormality. When these three factors were included in logistic regression models, independent associations with the latter two factors persisted. Inclusive or exclusive of “possible” cases of cerebral palsy, the ORs (with 95% CI limits) for dexamethasone treatment were 5.1 (1.3–21.1) and 13.4 (1.3–134.2), respectively.
The proportions of infants with head circumference, length, or weight below the 10th percentile for age were similar for the 2 groups. When analyzing data about the frequency of rehospitalization during the 1st year of life, a χ2 test for trend did not reach statistical significance (P = .1), although a higher proportion of placebo recipients were rehospitalized two or more times before 1 year adjusted age (P = .04). Excluding 2 dexamethasone-treated infants who were not evaluated at 1 year, a similar proportion of dexamethasone-treated infants and placebo recipients were alive and free from major neurosensory impairment at 1 year adjusted age (dexamethasone, 32/55 [58%] vs placebo, 37/61 [61%]).
The current study was undertaken to assess whether the beneficial effect of dexamethasone on pulmonary function, as indicated by a substantial number of randomized controlled trials,1–11leads to improved health and development after the neonatal period. Our analysis suggests that among infants surviving to 1 year of age, dexamethasone does not improve health or developmental outcome. On one hand, the higher prevalence of cerebral palsy at 1 year of age among dexamethasone-treated infants raises concern about the safety of dexamethasone. However, this finding must be viewed in the context of a strong trend toward improved survival among dexamethasone-treated infants. As described in an accompanying paper, 88% of dexamethasone-treated infants survived to 1 year adjusted age, compared with 74% of placebo recipients (P = .066; Fisher's exact test). Although not statistically significant, this mortality difference could have resulted in selection bias if the excess survival among dexamethasone-treated infants was experienced by infants whose a priori risks of adverse outcomes were higher than average. In addition, when one compares the proportion of infants who were alive and free from major neurosensory impairment at 1 year, no significant difference is found.
In contrast to observational studies,34 ,35 the results of this and other randomized trials2 ,3 ,7 ,9 ,10 ,36 ,37 do not indicate that the risk of visual impairment is increased by dexamethasone treatment. Although short-term effects of dexamethasone on growth are well-documented,38–40 our results and those of others11 39–40 suggest that these effects do not persist in later infancy. The trend, which we observed toward a lower rate of rehospitalization among dexamethasone-treated infants, is consistent with their lower rate of chronic lung disease, because this condition increases the risk of lower respiratory illness during infancy.21
Developmental follow-up has been reported after three previous randomized trials of dexamethasone. In the trial reported by Cummings and associates,2 a 42-day course, but not an 18-day course, of dexamethasone was associated with improved outcome at 18 months of age. However, the small sample used for this study (5 control subjects and 9 treated infants) limits the information that can be derived. The European Collaborative Dexamethasone Study involved a considerably larger sample of infants (209 infants); however, 40% of infants randomized to placebo were later treated with open dexamethasone, thereby attenuating potential associations between treatment and developmental outcome.5 ,14 The results reported here are consistent with those described by Yeh and colleagues, who found an increased risk of abnormal neuromotor function among infants given a 21-day course of dexamethasone beginning on the 1st day of life.15
Possible explanations for the higher rate of cerebral palsy among dexamethasone recipients, as observed here, include improved survival of infants with extant brain damage at study entry, improved survival of infants who were at increased susceptibility to causes of brain damage after randomization, and a direct adverse effect of dexamethasone on brain development. Experiments in which glucocorticoid treatment resulted in reduced cerebellar weight in 7-day-old rats lend additional support for the possibility that dexamethasone has a direct effect on brain development.41 It seems plausible that the slowing of growth among infants treated with dexamethasone is associated with impaired brain growth at a time when developing neural tissue is particularly vulnerable to injury. In support of this possibility is a recent report describing a decreased rate of head growth among infants treated with dexamethasone.42
In our study, and in a smaller study by Noble-Jamieson and co-workers,43 more dexamethasone-treated infants had major cranial ultrasound abnormalities, but this difference was not statistically significant in either study. Even if an association were assumed, our study could not establish causality, absent a prospective schedule for cranial ultrasonography. Nonetheless, the results of multivariate analysis suggest that an increased risk of major cranial abnormality does not explain fully the observed association between dexamethasone and cerebral palsy.
Despite the higher rate of cerebral palsy among dexamethasone-treated infants, these infants had similar scores on the Bayley Scales of Infant Development. One possible explanation is that because psychologists were not aware of infants' cranial ultrasound findings, their evaluations were not subject to the ascertainment bias, as could have affected neurologic assessments. Alternatively, the lack of effect of dexamethasone on Bayley scores may simply reflect the fact that 80% (12 of 15) of study infants with cerebral palsy had either spastic diplegia or hemiplegia. Prematurely born children with these forms of cerebral palsy typically have mild rather than severe developmental delay.44 ,45
Several possible limitations of our study should be noted. As discussed above, selection bias because of differential survival rates across the 2 study groups could explain the greater risk of cerebral palsy among dexamethasone recipients. Second, it is possible that clinicians, who were aware of infant's cranial ultrasound findings, introduced expectation bias into the diagnosis of cerebral palsy (greater expectation for infants with major cranial ultrasound examinations). It should be noted, however, that 4 dexamethasone-treated infants who were classified as having cerebral palsy did not have a major abnormality on cranial ultrasound examination. Third, the sample size chosen for our study was based on the objective of detecting what we considered to be a clinically significant effect of dexamethasone on the duration of ventilator dependency. Therefore, we may have failed to detect clinically significant effects on other outcomes, such as growth and mental development. A fourth possible limitation is related to the age at which infants were evaluated, because in some cases, signs of cerebral palsy that are present at 1 year may resolve during early childhood.46 Conversely, some experts have advised against ruling out cerebral palsy based on neurologic examinations performed at 1 year of age.47 More definitive information about the children described here may be forthcoming as we are currently evaluating them at 4 years of age. Finally, the generalizability of our study could be questioned if the dosages or schedules for dexamethasone therapy that are used by clinicians differ substantially from those used in our trial.
Despite these limitations, our findings have implications for clinical practice and future research. This is the only large trial of dexamethasone given after the first 2 weeks to infants treated routinely with surfactant, in which crossing over (of placebo infants to dexamethasone treatment) was not permitted, and it is one of only a small number of trials from which long-term data are available. In agreement with previous studies, our results indicate that dexamethasone reduces the risk of chronic lung disease, but does not improve long-term outcome. Together with other large trials,5 ,11 ,35 our study supports the use of dexamethasone in infants who are at risk for serious, potentially fatal, chronic lung disease. Because more conclusive information about the effects of dexamethasone on the brain is needed, this therapy should be considered experimental for infants who would not be predicted to develop severe lung disease. It follows, then, that derivation of methods for predicting severe chronic lung disease from attributes known during the first 1 to 2 weeks of life could prove useful when considering dexamethasone for a particular infant. For infants who would be predicted to develop mild, but not life-threatening, chronic lung disease, randomized, placebo-controlled trials, including long-term follow-up, would seem to be ethical. For infants who would be predicted to develop severe chronic lung disease, comparisons of alternative schedules of dosage and timing could provide additional information about the long-term effects of dexamethasone.
This research was supported by the North Carolina Department of Environment, Health, and Natural Resources, and the Brenner Children's Hospital, Winston-Salem, North Carolina.
- Received April 21, 1998.
- Accepted November 19, 1998.
Reprint requests to (T.M.O) Department of Pediatrics, Wake Forest University School of Medicine, Medical Center Blvd, Winston-Salem, NC 27157. E-mail:
Findings from this study were presented in part at the Annual Meeting of the American Pediatric Society/Society for Pediatric Research; May 6, 1997; Washington, DC.
- Avery GB,
- Fletcher AB,
- Kaplan M,
- Brudno DS
- Kazzi NJ,
- Brans YW,
- Poland RL
- Durand M,
- Sardesai S,
- McEvoy C
- Rastogi A,
- Akintorin SM,
- Bez ML,
- Morales P,
- Pildes RS
- ↵Yeh TF, Lin YJ, Hsieh WS, et al. Early postnatal dexamethasone therapy for the prevention of chronic lung disease in preterm infants with respiratory distress syndrome: a multicenter clinical trial. Pediatrics 1997;100(4). URL: http://www.pediatrics.org/cgi/content/full/100/4/e3
- Jones R,
- Wincott E,
- Elbourne D,
- Grant A
- ↵Yeh TF, Lin YJ, Huang CC, Chen YJ, Tsai WF, Lien YJ. Early postnatal (<12 hrs) dexamethasone therapy for prevention of BPD in preterm infants with RDS—a two-year follow-up study. Pediatrics. 1998;101(5). URL: http://www.pediatrics.org/cgi/content/full/101/5/e7
- Committee on An International Classification of Retinopathy of Prematurity
- O'Shea TM,
- Klinepeter KL,
- Goldstein DJ,
- Jackson BW,
- Dillard RG
- ↵Bayley N. The Bayley Scales of Infant Development. New York, NY: Psychological Corporation; 1969
- ↵Sparrow SS, Balla DA, Cicchetti DV. Vineland Adaptive Behavior Scales. Circle Pines, MN: American Guidance Service, Inc; 1984
- ↵Bayley N. Bayley Scales of Infant Development. 2nd ed. San Antonio, TX: Psychological Corporation; 1993
- Shennan AT,
- Dunn MS,
- Ohlsson A,
- Lennox K,
- Hoskins EM
- Hyde I,
- English RE,
- Williams JD
- Lubchenco LO,
- Hansman C,
- Boyd E
- ↵Fleiss JL. Statistical Methods for Rates and Proportions. 2nd ed. New York, NY: John Wiley and Sons; 1981
- ↵PC-SAS, Cary, NC: SAS Institute
- ↵STATXACT, Cambridge, MA: Cytel
- Batton DG,
- Roberts C,
- Trese M,
- Maisels MJ
- Brownlee KG,
- Ng PC,
- Henderson MJ,
- Smith M,
- Green JH,
- Dear PRF
- Gibson AT,
- Pearse RG,
- Wales JKH
- ↵Jarvis S, Hey E. Measuring disability and handicap to cerebral palsy. In: Stanley F, Alberman E, eds. The Epidemiology of Cerebral Palsies. Philadelphia, PA: JB Lippincott Co; 1984;41–42
- ↵Paneth N, Rudelli R, Kazam E, Monte W. Prognosis. In: Brain Damage in the Preterm Infant. London, UK: MacKeith Press; 1994:171–185
- Nelson KB,
- Ellenberg JH
- Hamill PVV,
- Drizd TA,
- Johnson CL,
- Reed RB,
- Roche AF,
- Moore WM
- Copyright © 1999 American Academy of Pediatrics