Prophylaxis of Early Adrenal Insufficiency to Prevent Bronchopulmonary Dysplasia: A Multicenter Trial
Background. Infants developing bronchopulmonary dysplasia (BPD) show decreased cortisol response to adrenocorticotropic hormone. A pilot study of low-dose hydrocortisone therapy for prophylaxis of early adrenal insufficiency showed improved survival without BPD at 36 weeks’ postmenstrual age, particularly in infants exposed to histologic chorioamnionitis.
Methods. Mechanically ventilated infants with birth weights of 500 to 999 g were enrolled into this multicenter, randomized, masked trial between 12 and 48 hours of life. Patients received placebo or hydrocortisone, 1 mg/kg per day for 12 days, then 0.5 mg/kg per day for 3 days. BPD at 36 weeks’ postmenstrual age was defined clinically (receiving supplemental oxygen) and physiologically (supplemental oxygen required for O2 saturation ≥90%).
Results. Patient enrollment was stopped at 360 patients because of an increase in spontaneous gastrointestinal perforation in the hydrocortisone-treated group. Survival without BPD was similar, defined clinically or physiologically, as were mortality, head circumference, and weight at 36 weeks. For patients exposed to histologic chorioamnionitis (n = 149), hydrocortisone treatment significantly decreased mortality and increased survival without BPD, defined clinically or physiologically. After treatment, cortisol values and response to adrenocorticotropic hormone were similar between groups. Hydrocortisone-treated infants receiving indomethacin had more gastrointestinal perforations than placebo-treated infants receiving indomethacin, suggesting an interactive effect.
Conclusions. Prophylaxis of early adrenal insufficiency did not improve survival without BPD in the overall study population; however, treatment of chorioamnionitis-exposed infants significantly decreased mortality and improved survival without BPD. Low-dose hydrocortisone therapy did not suppress adrenal function or compromise short-term growth. The combination of indomethacin and hydrocortisone should be avoided.
- adrenal insufficiency
- bronchopulmonary dysplasia
- adrenocorticotropic hormone
- extremely low birth weight infants
Relative adrenal insufficiency in the face of acute stress or illness is a well-recognized phenomenon.1,2 Affected individuals have apparently normal adrenal function under most circumstances but are unable to respond adequately to severe stress, displaying cardiovascular instability and increased mortality.1–3 A recent randomized trial showed that low-dose hydrocortisone therapy reduced mortality in adults with septic shock and relative adrenal insufficiency.4
Extremely premature infants may be at increased risk for a similar phenomenon during acute illness because of developmental immaturity of the hypothalamic-pituitary-adrenal axis. As reviewed by Mesiano and Jaffe,5 during much of gestation the fetal adrenal gland is deficient in the enzyme 3β-hydroxysteroid dehydrogenase (3βHSD), which catalyzes an essential step in the production of cortisol from cholesterol. After the embryonic period, the human fetal adrenal cortex probably does not produce cortisol de novo before 23 weeks of gestation, and in the undisturbed fetus may not do so until as late as 30 weeks of gestation. Instead, the fetus synthesizes cortisol from placental progesterone, bypassing the need for 3βHSD. After delivery, placental progesterone is no longer available, and limited 3βHSD may reduce the extremely premature infant’s ability to produce cortisol when needed during illness. Lower cortisol values and decreased cortisol response to adrenal stimulation have been reported in premature infants who have higher illness-severity scores, those who are mechanically ventilated, and those receiving vasopressor support for hypotension.6–8 Low cortisol values have been associated also with increased mortality in this population.9
Adrenal insufficiency is also associated with amplified inflammatory responses, because cortisol is essential for resolution of inflammation10,11; indeed, this is one of cortisol’s primary physiologic actions. Infants who develop bronchopulmonary dysplasia (BPD) have been shown to have decreased adrenal function early in life as well as increased indicators of both prenatal and postnatal inflammation.6,12–19 We proposed that the increased inflammation seen in infants developing BPD may in part reflect inadequate adrenal function in the face of acute illness and that low-dose cortisol replacement therapy might decrease the incidence of BPD and improve clinical stability in the immediate newborn period. To test this hypothesis, a randomized pilot study was conducted that showed significant improvement in survival without BPD in the infants treated with hydrocortisone.20
In previous studies of adrenal function in extremely premature infants, we did not find differences in cortisol concentrations in the first 48 hours of life in infants who subsequently developed BPD, perhaps because of the surge in cortisol production during labor and delivery and/or delayed metabolism of passively transmitted maternal cortisol. However, we have found lower cortisol values as early as the third day of life in infants with patent ductus arteriosus and in those who develop BPD.12,13 Because we wished to prevent this fall in cortisol concentrations in hopes of preventing these adverse events, we designed this multicenter trial as a prophylaxis study, enrolling intubated, extremely low birth weight infants as a group at very high risk for the primary adverse outcome.21–23
This trial differed from previous trials of postnatal dexamethasone therapy to prevent BPD in premature infants in (1) the hypothesis that preventing the development of adrenal insufficiency of prematurity would improve clinical outcomes; (2) the dose of glucocorticoid, significantly lower than in previous trials, intended to mimic cortisol concentrations seen in patients under stress1–4; and (3) the glucocorticoid used, hydrocortisone, which is metabolized to the endogenous glucocorticoid, cortisol.
Population and Randomization
All infants between 500 and 999 g birth weight admitted to participating units were screened for eligibility. Within this weight group, infants were eligible if they were mechanically ventilated at study entry, between 12 and 48 hours of life. Exclusion criteria included (1) major congenital anomaly, (2) congenital sepsis, (3) postnatal glucocorticoid treatment other than hydrocortisone, or (4) triplet or higher-order multiple gestation. This protocol was approved by institutional review boards at all participating institutions; parental consent was obtained before enrollment.
Randomization was stratified by study center and birth weight (500–749 and 750–999 g) using a permuted-blocks scheme with blocks of 6 within each stratum. Randomization lists were provided to each pharmacy in a sealed envelope by the Data Coordinating Center at Pennsylvania State University. Only the pharmacists preparing the drug were aware of the group assignment. All other personnel were masked. Twins were randomized together to the same study arm.
Infants received an equal volume of normal saline placebo or low-dose hydrocortisone sodium succinate (Solu-Cortef Plain, Pharmacia and Upjohn, Kalamazoo, MI), 1 mg/kg per day (∼8–10 mg/m2 per day) divided twice a day for 12 days followed by 0.5 mg/kg per day for 3 days. Previous studies showed that this dose resulted in plasma cortisol concentrations similar to physiologic concentrations in patients under stress.1–4,20,24
Cortisol values were obtained at study entry and again at ∼1 week of age. An adrenocorticotropic hormone (ACTH) challenge was administered between 48 hours after the last dose of study drug and 28 days of life. Sampling was done before and 30 minutes after administration of cosyntropin (Cortrosyn, Amphastar, Rancho Cucamonga, CA). To compare dose response, 1.0 μg/kg of cosyntropin was given to the first 150 infants, and 0.1 μg/kg was given to all subsequent infants. Infants were not tested while receiving open-label glucocorticoid. Specimens were analyzed in duplicate by radioimmunoassay (Diagnostic Products Corp, Los Angeles, CA) at the Tufts-New England Medical Center General Clinical Research Center.
Because the pilot study demonstrated particular benefit for infants exposed to chorioamnionitis, this group was of specific a priori interest.20 All placental histology was therefore reviewed by central readers (N. Joste, MD, and M. Wills, MD; see the Appendix). Chorioamnionitis was defined as the presence of acute inflammatory cells in the subchorion, chorion, or amnion. The presence of fetal vasculitis (fetal inflammatory response) was recorded for the umbilical vessel walls, Wharton’s jelly, and the chorionic plate. Inflammation at all sites was graded as mild, moderate, or severe.25 Because the incidence of cystic periventricular leukomalacia may be increased after higher-dose dexamethasone treatment,21 cranial ultrasounds performed after 4 weeks of age were reviewed for periventricular leukomalacia by a central reader (D.K.B. Boal, MD).
Study Monitoring and Outcomes
The trial was monitored by a data safety monitoring committee. This committee was notified of specific adverse events (death, sepsis, and gastrointestinal perforation) within 72 hours and reviewed other adverse outcomes approximately quarterly. Although hydrocortisone therapy had not previously been associated with spontaneous gastrointestinal perforation, this adverse event was monitored closely because of previous reports linking higher-dose dexamethasone therapy to spontaneous perforation.22
The primary outcome measure was survival without BPD, defined as oxygen dependence at 36 weeks’ postmenstrual age (PMA). Physiologic BPD was also evaluated in infants who remained in a study-site hospital at 36 weeks’ PMA, as follows. Infants receiving oxygen were placed in a hood for 4 hours, and the lowest fraction of inspired oxygen (Fio2) required to maintain an oxygen saturation of ≥90% for 30 minutes of quiet wakefulness or sleep was recorded; adjustment for altitude was made in Denver, Colorado, and Albuquerque, New Mexico, by considering an Fio2 of 0.25 equal to room air at sea level.
Secondary outcomes included death before 36 weeks’ PMA, death before discharge, BPD in survivors, duration of mechanical ventilation, oxygen therapy and hospital stay, and weight and head circumference at 36 weeks’ PMA. Daily information recorded for the first 28 days included respiratory support, blood pressure, electrolyte and glucose data, fluid intake, enteral intake, cultures done, medications received, and evidence of gastrointestinal bleeding. At 36 weeks’ PMA, ongoing respiratory support and medications were recorded, as were cumulative adverse events including patent ductus arteriosus, respiratory complications, nosocomial infection, necrotizing enterocolitis (NEC), gastrointestinal perforation, intracranial hemorrhage, periventricular leukomalacia, retinopathy of prematurity, and open-label steroid therapy. All clinical data were entered into the database twice for quality assurance.
Based on an anticipated incidence of survival without BPD at 36 weeks’ PMA of 35%,22,23 a sample size of 712 births was required to achieve a power of 0.80 to detect a difference of 10 percentage points in the primary outcome with α = .05. Including eligible second twins, the anticipated sample size was 790 infants. Because the final sample size was 360 infants, statistically insignificant results must be viewed with caution because of the increased probability of a type II error.
All randomized infants were included in an intent-to-treat analysis. Baseline comparisons used 2-sample t tests or Wilcoxon-Mann-Whitney tests for continuous outcomes and χ2 or Fisher’s exact tests for categorical outcomes. Outcomes and adverse events were analyzed by using analysis of variance or covariance for continuous outcomes and logistic regression or Fisher’s exact tests for binary outcomes. Area under the curve for continuous daily clinical outcomes was computed for infants alive and enrolled at the end of treatment and compared by analysis of variance. For binary daily outcomes, the percent of treatment days exposed was compared by using a generalized binomial model.26 Comparisons of length of therapy (eg, oxygen, length of stay) were made using Cox proportional hazards models.27 Cortisol values were log-transformed for analyses and are presented as nanomoles per liter (÷27.6 = μg/dL). Where possible, the impact of correlation between infant responses within twin births was investigated by using (1) a linear mixed-effects model extension of the 2-sample t test, analysis of variance, or analysis of covariance for continuous outcomes and (2) generalized estimating equations with a logit link in place of χ2 tests or logistic regression models for binary outcomes.28,29 Because these analyses showed minimal impact, results presented are based on analyses that assume patients are independent.
One thousand twenty-five infants were screened, and 360 patients were enrolled between November 1, 2001, and April 30, 2003 (Fig 1). Because of an increased incidence of apparently spontaneous gastrointestinal perforation in the hydrocortisone group, the study was stopped at the recommendation of the data safety monitoring committee. Three randomized infants were later found to have major congenital anomalies (trisomy 21, tetralogy of Fallot, ileal atresia), and 2 had congenital sepsis; all were included in the intent-to-treat analyses. Patient enrollment at the 9 study centers ranged from 16 to 84 patients per center (mean: 40; median: 34).
Table 1 shows population characteristics. The hydrocortisone-treated group contained significantly more outborn infants. Several other factors marginally favored the placebo group: fewer males, fewer white infants, fewer vaginal deliveries; more prenatal steroids; and less vasopressor and glucocorticoid support at study entry. Table 2 shows the primary outcome, first for the entire study population and then separated by exposure to histologic chorioamnionitis. Primary outcomes were similar for the overall study population; however, for infants exposed to chorioamnionitis, the hydrocortisone-treated infants had significantly improved survival without BPD and a significantly lower mortality rate than placebo-treated infants. Multivariate analysis including only infants with known placental histology revealed a significant interaction between chorioamnionitis and treatment group (P = .025). In this analysis, the odds ratio (OR) for survival without BPD in chorioamnionitis-exposed infants favored the hydrocortisone-treated infants: 2.84 (95% confidence interval [CI]: 1.21–6.67). For infants without chorioamnionitis, the OR was 0.72 (95% CI: 0.31–1.65).
Infants exposed to chorioamnionitis were then analyzed with regard to the presence or absence of fetal inflammation. Ninety-nine infants (67%) had fetal inflammation. For chorioamnionitis-exposed infants without fetal inflammation, survival without BPD was not different between groups (P = .43). For infants with fetal inflammation, hydrocortisone-treated infants had significantly improved survival without BPD (P = .01; OR: 1.45–17.96).
During treatment, the occurrence of hyponatremia (sodium < 130 mEq/L), hypernatremia (sodium > 150 mEq/L), hyperkalemia (central potassium > 7.0 mEq/L), hyperglycemia (>180 mM/L), hypertension (mean blood pressure > 50 mm Hg or systolic blood pressure > 70 mm Hg), and upper or lower gastrointestinal bleeding were similar between groups. Seventy-four infants in the hydrocortisone group (41%) were treated with insulin, vs 62 (34%) in the placebo group (P = .192), for a median (25th–75th percentile) of 3 (2–5) days in each group. Four hydrocortisone-treated infants and 5 placebo infants received antihypertensive therapy. Total daily fluid intake, enteral intake, and urine output were similar between groups, as was vasopressor support. Serum sodium and mean arterial blood pressure were significantly higher in hydrocortisone-treated infants (P < .001 and P = .022, respectively).
At study entry, cortisol values were similar between study groups: median (25th–75th percentile) = 408 (246–802) nmol/L for hydrocortisone-treated infants and 471 (243–910) nmol/L for placebo infants. As anticipated,31 infants exposed to chorioamnionitis had higher cortisol values than those not exposed: 474 (269–944) vs 360 (184–673) nmol/L (P < .001). At the end of the first week of life, excluding infants receiving open-label steroids, cortisol values in the hydrocortisone-treated infants were 508 (335–1114) vs 360 (241–514) nmol/L in the placebo group (P < .001). After treatment, there was no difference between treatment groups in basal or stimulated cortisol values. Hydrocortisone-treated infants had a median basal cortisol value of 306 (219–433) vs 303 (216–405) nmol/L for the placebo group. After 0.1 μg/kg cosyntropin, cortisol values in treated infants were 476 (326–630) vs 428 (325–517) nmol/L for the placebo group. After 1.0 μg/kg cosyntropin, treated infants had cortisol concentrations of 651 (554–941) vs 706 (592–922) nmol/L for placebo infants. The cosyntropin dose of 0.1 μg/kg did not discriminate between clinical variables previously shown to affect response to ACTH.32 Adjusting for gestational age and chorioamnionitis, infants developing BPD had a lower response to 1.0 μg/kg cosyntropin than those recovering without BPD (P = .04).
Table 3 summarizes other clinical outcomes. Notably, neither weight nor head circumference were decreased in the hydrocortisone-treated infants. Fewer hydrocortisone-treated infants received open-label steroids during the treatment period, but the total number of infants receiving open-label steroids during hospitalization was not different between treatment groups. Restricting the analysis of other outcomes to those infants with chorioamnionitis (n = 149) did not move any of these outcomes across the significance threshold of P = .05, with the exception of “systemic steroids during treatment” (18% vs 28%; P = .28), “retinopathy of prematurity any stage” (82% vs 90%; P = .02), and “gastrointestinal perforation spontaneous or NEC-related” (14% vs 3%; P = .02).
The incidence of gastrointestinal perforation considered to be spontaneous in the best judgment of the principle investigator at the study site occurred significantly more often in the hydrocortisone-treated infants (P = .01). Of the 21 affected infants, 3 in each treatment group died before discharge. Hydrocortisone-treated infants who received indomethacin were more likely to experience apparently spontaneous gastrointestinal perforation than placebo-treated infants who received indomethacin (P < .001; Table 4), suggesting an interactive effect. The incidence of perforation in hydrocortisone-treated infants who did not receive indomethacin was similar to that of the placebo group. No relationship was seen between prenatal indomethacin exposure and perforation.
In this study, 55% of infants received prophylactic indomethacin, and 63% received indomethacin during the first 2 days of life. Table 4 separates infants first by assigned treatment group and indomethacin exposure and then by any systemic glucocorticoid exposure, any indomethacin exposure, both or neither, before gastrointestinal perforation during the treatment period. Infants experiencing apparently spontaneous gastrointestinal perforation were more likely to be male (81% vs 50% of the overall population; P = .006) and outborn (33% vs 11%; P = .007), with lower gestational age (P = .011). The incidence in infants exposed to chorioamnionitis was 6% (9 of 149). Of infants delivered for maternal indication (n = 55), only 1 had a spontaneous perforation (2%). Excluding infants receiving open-label glucocorticoid therapy, infants with spontaneous perforation had significantly higher cortisol values before study entry (median [25th–75th percentile]: 1368 [496–2984] vs 420 [240–819] nmol/L or 50 [18–108] vs 15 [9–30] μg/dL; P < .001). These infants continued to have higher cortisol values at the end of the first week of life (placebo group: 784 [546–1003] vs 356 [239–495] nmol/L or 28 [20–36] vs 13 [9–18] μg/dL; P = .036; treatment group: 817 [456–1457] vs 497 [332–1024] nmol/L or 30 [17–53] vs 18 [12–37] μg/dL; P = .039).
In contrast to the pilot study on which it was based, this multicenter trial did not show improved survival without BPD at 36 weeks’ PMA after early, low-dose hydrocortisone therapy for prophylaxis of adrenal insufficiency. This held true regardless of whether BPD was defined clinically or physiologically. However, in infants exposed to histologic chorioamnionitis, survival without BPD was significantly greater in the hydrocortisone-treated infants than in the placebo group, and mortality was significantly lower. This benefit appeared to be specific to those infants with evidence of fetal inflammatory response.
The overall incidence of survival without clinical BPD in this population is consistent with previous reports.22,23 The increased risk for development of BPD in placebo infants exposed to chorioamnionitis is consistent with both the pilot study and several previous reports linking exposure to prenatal inflammation with increased risk for BPD.15–18,20,33 The relationship of histologic chorioamnionitis to lung disease in the preterm infant is complex. Increased prenatal inflammation promotes lung maturation, resulting in decreased severity of acute lung disease.31,34,35 However, when postnatal inflammatory insults such as mechanical ventilation and infection are added, prenatal inflammation seems to amplify lung injury, promoting the development of BPD.33,35 Because cortisol is essential for resolving inflammation, relative adrenal insufficiency may contribute to BPD by promoting increased lung inflammation.
There is no accepted normal range for cortisol concentrations in the preterm infant. Clinically well preterm infants may have quite low cortisol concentrations without evidence of harm.36 However, lower cortisol concentrations and/or decreased response to adrenal stimulation have been documented in sicker infants, in those receiving vasopressor support or mechanical ventilation, and in those developing BPD or dying.6–9,12,13 These lower values are contrary to the expectation that the seriously ill should have higher cortisol concentrations than well individuals and may be similar to the relative adrenal insufficiency documented in other critically ill patients, which manifests as cardiovascular instability and increased mortality.1–3 The decreased mortality rate in hydrocortisone-treated infants born after chorioamnionitis may be analogous to the decreased mortality reported in adults with septic shock and adrenal insufficiency treated with hydrocortisone.4 Hydrocortisone-treated infants in this study also showed higher blood pressure without an increase in hypertension.
We found that infants developing BPD had decreased cortisol response to ACTH stimulation after the treatment period, which is consistent with several previous studies showing decreased basal and/or stimulated cortisol values early in life in such infants6,12,14,20,37 and with the hypothesis that decreased adrenal function promotes the development of BPD in intubated extremely low birth weight infants. In 1 previous study, Ng et al38 reported that although lower cortisol values in the first week of life correlated with a longer duration of oxygen therapy, by the second week the reverse was true: higher cortisol values correlated with longer duration of oxygen therapy, perhaps indicating a stress response to continuing illness. The infants in our study were smaller and sicker, and all were intubated at study entry. These more immature infants might be at risk for a longer period of developmental adrenal immaturity. In our study, we also included the presence or absence of histologic chorioamnionitis in the analyses. Because chorioamnionitis increases both inflammation and cortisol concentrations in the fetus and newborn,15,18,20,31,34,35 not including this factor in the analysis may mask differences between groups.
Hydrocortisone treatment resulted in serum concentrations similar to values previously documented in other patient populations under stress.1–4 However, a few infants had substantially elevated cortisol concentrations, and those who experienced spontaneous gastrointestinal perforation had significantly higher values than those who did not, both before and during therapy, suggesting that cortisol concentrations should be monitored in infants being treated with hydrocortisone.
Patient enrollment in this study was stopped because of a significant increase in apparently spontaneous perforation of the gastrointestinal tract in hydrocortisone-treated infants. Differentiating this entity from NEC can be difficult; however, there was no difference in NEC-associated perforation between groups. Spontaneous gastrointestinal perforation occurred less often in this placebo group (2%) than has been previously reported in such infants: ∼4% overall in this birth weight group39 and 7% in populations similar to this study.21,22 There seemed to be an interaction between hydrocortisone and indomethacin therapy, as previously reported with higher-dose dexamethasone given in the first days of life22; however, indomethacin therapy was not randomized in either trial. Evaluating the interaction is also made more difficult because almost two thirds of these patients received indomethacin in the first 2 days of life. Other reported experience with low-dose hydrocortisone therapy in such infants is quite limited; however, in the absence of early indomethacin, low-dose hydrocortisone therapy administered as described in this study has not previously been associated with an increased incidence of spontaneous gastrointestinal perforation20 (1.5% of 199 treated infants [T. Lacaze, MD, PhD, written communication, 2003]).
Spontaneous gastrointestinal perforation in the extremely premature infant has been associated with early indomethacin therapy and with higher-dose dexamethasone therapy given early in life21,22,40; multiple mechanisms may participate in its pathogenesis. Indomethacin decreases blood flow to the intestine and causes direct mucosal injury, amplified by its enterohepatic circulation.41,42 In experimental models, indomethacin also inhibited prostaglandin synthesis while increasing intestinal contractility, myeloperoxidase activity, inducible nitric oxide synthase activity, and bacterial numbers in the mucosa, all of which may contribute to mucosal damage.43
Physiologic concentrations of glucocorticoid protected against indomethacin-induced damage in an adult model of indomethacin-induced injury.44 In the extremely preterm infant, however, glucocorticoids may contribute to perforation by accelerating normal intestinal mucosal development. Gordon et al45,46 reported that administration of dexamethasone to neonatal mice altered local concentrations of growth factors in the intestine, producing mucosal hypertrophy and smooth muscle thinning similar to that seen in infants with spontaneous perforation. In our study, the significantly higher cortisol concentrations seen in infants developing spontaneous gastrointestinal perforation are consistent with such a dose-dependent effect.
With the notable exception of spontaneous gastrointestinal perforation, other adverse effects previously reported with higher-dose dexamethasone therapy were not seen.21,22 Specifically, there was no evidence of suppression of either weight or head growth at 36 weeks’ PMA, no increase in hypertension or treatment for hypertension, no increase in sepsis (including fungal sepsis), and no increase in periventricular leukomalacia on late cranial ultrasound. When the adrenal axis was tested after therapy, the hydrocortisone-treated infants showed no suppression of either basal or stimulated cortisol values. Somatic and neurodevelopmental outcomes will be evaluated when the infants are 18 to 22 months’ corrected age.
The approach taken in this study, early prophylactic hydrocortisone treatment to prevent the development of adrenal insufficiency in infants at highest risk of developing BPD, failed to benefit its target group. However, the significant improvement seen in infants born after chorioamnionitis (those anticipated to benefit most from the treatment) warrants additional exploration of how low-dose hydrocortisone therapy might be used in extremely preterm infants to provide benefit without increasing risk.
Participating Institutions, Newborn Intensive Care Units, and Other Investigators
Pennsylvania Hospital: Pennsylvania Hospital Newborn Intensive Care Unit (Soraya Abbasi, MD, and Toni Mancini, RN).
Tufts University: New England Medical Center Newborn Intensive Care Unit (Brenda MacKinnon, RNC); Tufts University General Clinical Research Center (Roland Stewart, MS).
University of Colorado: University of Colorado Hospital Newborn Intensive Care Unit (Susan Townsend, MD, and Shannon Collins, RN); Children’s Hospital of Denver Newborn Intensive Care Unit; Denver Health Medical Center Newborn Intensive Care Unit (Owen P. O’Meara, MD); and Exempla-St Joseph’s Hospital Newborn Intensive Care Unit.
Johns Hopkins University: Johns Hopkins University Newborn Intensive Care Unit (Pamela Donohue, ScD, and Jennifer Shepard, RN); Johns Hopkins Bayview Medical Center Newborn Intensive Care Unit (Maureen Gilmore, MD); and Greater Baltimore Medical Center Newborn Intensive Care Unit (Maria Pane, MD).
St Joseph Regional Medical Center: St Joseph Regional Medical Center Newborn Intensive Care Unit (Colleen Alex, RN).
Virginia Commonwealth University: Virginia Commonwealth University Medical Center Newborn Intensive Care Unit (Gail Barker, RN).
State University of New York at Buffalo: Children’s Hospital of Buffalo Newborn Intensive Care Unit (Kirsten Blessing-Hanagan, RN).
Pennsylvania State University: Department of Health Evaluation Sciences (Lisa Szwejbka, MPH, and Jennifer Lucier, BS); and Department of Radiology (Danielle K.B. Boal, MD).
University of New Mexico: Children’s Hospital of New Mexico Newborn Intensive Care Unit (Rebecca Montman, RN); and University of New Mexico Department of Pathology (Nancy Joste, MD, and Marcia Wills, MD).
Children’s Hospitals and Clinics of Minneapolis/St Paul: Children’s Hospital of Minneapolis Newborn Intensive Care Unit (Molly Maxwell, RN); and Children’s Hospital of St Paul Newborn Intensive Care Unit (Pat Meyers, RN).
This work was supported by National Institute of Child Health and Human Development grant R01-HD38540, grant MO1 RROOO54 from the General Clinical Research Centers Programs at the University of New Mexico, Tufts-New England Medical Center grant 5MO1 RROO997, and University of Colorado grant MO1-RROOO69.
We are indebted to Drs Michael O’Shea, Avroy Fanaroff, John Johnson, Nigel Paneth, and Linda Wright for service as members of the Data Safety Monitoring Committee; Dr LuAnn Papile for manuscript review; the coordinators and staffs of the participating institutions; and the patients and families who participated in this study.
- Accepted September 1, 2004.
- Reprint requests to (K.L.W.) Department of Pediatrics/Neonatology, MSC10 5590, 1 University of New Mexico, Albuquerque, NM 87131-0001. E-mail:
No conflict of interest declared.
- ↵Ng PC, Lee CH, Kam CWK, et al. Transient adrenocortical insufficiency of prematurity and systemic hypotension in very low birth weight infants. Arch Dis Child Fetal Neonatal Ed.2004;89 :F119– F126
- ↵Watterberg KL, Scott SM. Evidence of early adrenal insufficiency in babies who develop bronchopulmonary dysplasia. Pediatrics.1995;95 :120– 125
- ↵Watterberg KL, Scott SM, Backstrom C, Gifford KL, Cook KL. Links between early adrenal function and respiratory outcome in preterm infants: airway inflammation and patent ductus arteriosus. Pediatrics.2000;105 :320– 324
- ↵Watterberg KL, Demers LM, Scott SM, Murphy S. Chorioamnionitis and early lung inflammation in babies who develop bronchopulmonary dysplasia. Pediatrics.1996;97 :210– 215
- ↵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
- ↵The Vermont Oxford Network Steroid Study Group. Early postnatal dexamethasone therapy for the prevention of chronic lung disease. Pediatrics.2001;108 :741– 748
- ↵Watterberg KL, Cook K, Gifford KL. Pharmacokinetics of hydrocortisone in extremely low birth weight infants in the first week of life [abstract]. Pediatr Res.1996;39 :251A
- ↵McCullagh P, Nelder JA. Binary data. In: Generalized Linear Models. 2nd ed. London, England: Chapman and Hall; 1989:124–128
- ↵Cox DR. Regression models and life-tables (with discussion). J R Stat Soc [Ser B].1972;34 :187– 220
- ↵Liang K-Y, Zeger SL. Longitudinal data analysis using generalized linear models. Biometrika.1986;73 :13– 22
- ↵Watterberg KL, Scott SM, Naeye RL. Chorioamnionitis, cortisol, and acute lung disease in very low birth weight infants. Pediatrics.1997;99(2) . Available at: www.pediatrics.org/cgi/content/full/99/2/e6
- ↵Watterberg KL, the PROPHET study group. A cosyntropin dose of 0.1μg/kg is inadequate to test adrenal function in extremely low birth weight infants [abstract]. Pediatr Res.2004;55 :154A
- ↵Shimoya K, Taniguchi T, Matsuzaki N, et al. Chorioamnionitis decreased incidence of respiratory distress syndrome by elevating fetal interleukin-6 serum concentration. Hum Reprod.2000;15 :2234– 2240
- ↵Banks BA, Stouffer N Cnaan A, et al. Association of plasma cortisol and chronic lung disease in preterm infants. Pediatrics.2001;107 :494– 498
- ↵Ng PC, Lee CH, Lam CWK, et al. Early pituitary-adrenal response and respiratory outcomes in preterm infants. Arch Dis Child Fetal Neonatal Ed.2004;89 :F127– F130
- ↵Horbar JD, Carpenter JH, eds. Vermont Oxford Network 2002 Database Summary. Burlington, VT: Vermont Oxford Network; 2003
- ↵Rainsford KD, Stetsko PI, Sirko SP, Debski S. Gastrointestinal mucosal injury following repeated daily oral administration of conventional formulations of indomethacin and other non-steroidal anti-inflammatory drugs to pigs: a model for human gastrointestinal disease. J Pharm Pharmacol.2003;55 :661– 668
- ↵Tanaka A, Hase S, Miyazawa T, Ohno R, Takeuchi K. Role of cyclooxygenase (COX)-1 and COX2 inhibition in nonsteroidal anti-inflammatory drug-induced intestinal damage in rats: relation to various pathogenic events. J Pharmacol Exp Ther.2002;303 :1248– 1254
- ↵Filaretova L, Tanaka A, Miyazawa T, Kato S, Takeuchi K. Mechanisms by which endogenous glucocorticoid protects against indomethacin-induced gastric injury in rats. Am J Physiol Gastrointest Liver Physiol.2002;283 :G1082– G1089
- Copyright © 2004 by the American Academy of Pediatrics