Prolonged Indomethacin Exposure Is Associated With Decreased White Matter Injury Detected With Magnetic Resonance Imaging in Premature Newborns at 24 to 28 Weeks' Gestation at Birth
OBJECTIVES. Newborns delivered before 28 weeks' gestation commonly have white matter lesions on MRI that are associated with adverse neurodevelopmental outcomes. Our objective was to determine the risk factors for MRI-detectable white matter injury in infants delivered before 28 weeks' gestation who were treated with prophylactic indomethacin.
METHODS. This was a prospective cohort study conducted at the intensive care nursery at University of California San Francisco Children's Hospital. Patients included 57 premature newborns between 24 and 27 (+6 days) weeks' gestation at birth (October 1998 to October 2004). We identified perinatal and neonatal risk factors associated with moderate-severe “white matter injuries” (T1 signal abnormalities >2 mm or >3 areas of T1 abnormality) and moderate-severe “brain abnormality” (moderate-severe white matter injuries, any degree of ventriculomegaly, or severe intraventricular hemorrhage) on MRI. Infants were studied with MRI at 31.1 weeks’ postmenstrual age (median).
RESULTS. Moderate-severe white matter injuries were detected in 12 (21%) of 53 preterm newborns, and 20 (35%) of 57 had moderate-severe brain abnormality. Prolonged indomethacin exposure was the only risk factor independently associated with a lower risk of white matter injury or brain abnormality, even when adjusting for the presence of a hemodynamically significant PDA, gestational age at birth, prenatal betamethasone, systemic infection, and days of mechanical ventilation.
CONCLUSIONS. In this observational study, a longer duration of indomethacin exposure was associated with less white matter injury in infants delivered before 28 weeks' gestation. A randomized trial of prolonged indomethacin treatment is needed to determine whether indomethacin can decrease white matter injury and neurodevelopmental abnormalities.
Neurodevelopmental abnormalities are common after premature birth.1 These motor and cognitive deficits are not necessarily associated with the appearance of cystic periventricular leukomalacia or severe intraventricular hemorrhage (IVH) on ultrasound during the early neonatal period.1,2 Recent studies have shown that noncystic white matter injury is the characteristic pattern of brain injury in the premature newborn.3,4 Unlike cystic periventricular leukomalacia, noncystic white matter lesions on MRI are not well detected with ultrasound.4–7 These MRI-demonstrable noncystic white matter lesions occur before 40 weeks' postmenstrual age (PMA). In fact, in premature newborns studied serially, white matter injury is most readily detected early in life and becomes less apparent by term-equivalent age.3 Recent data suggest that noncystic white matter injury on MRI is associated with abnormalities of early motor and cognitive function.3
Unfortunately, in a recent prospective cohort study of preterm newborns (delivered between 24 and 34 weeks' PMA), we were unable to find any clinical risk factors that might identify noncystic white matter injury.3 In particular, risk factors associated previously with cystic PVL, such as clinical chorioamnionitis, sepsis, chronic lung disease, severe systemic hypotension, and necrotizing enterocolitis, were not associated with noncystic white matter injury on MRI. Given the heterogeneous nature of the group of infants that we examined previously, we sought to prospectively examine the association between perinatal and neonatal risk factors and the occurrence of MRI-detectable white matter injury in infants who were less diverse and more similarly treated.
At our center, all of the newborns delivered before 28 weeks' PMA are treated with a prophylactic course of indomethacin to prevent the development of a hemodynamically significant, symptomatic patent ductus arteriosus (PDA). PDA is a major risk factor for morbidity and mortality in premature newborns.8,9 Indomethacin, a prostaglandin synthesis inhibitor, can be used to pharmacologically close a symptomatic PDA.8,9 The prophylactic use of indomethacin in the first 3 days of life reduces the occurrence of a clinically symptomatic PDA and the rates of severe IVH in the premature newborn.9–12 In addition, a recent meta-analysis of 5 trials that examined the relationship between prophylactic indomethacin and the ultrasound diagnosis of periventricular leukomalacia or ischemic changes suggests that prophylactic indomethacin may prevent these ultrasound abnormalities.13 Because the rate of ductus reopening remains unacceptably high with a short course of prophylactic indomethacin treatment (≤3 days), increased attention is being paid to a more prolonged course of indomethacin (≥4 days).14,15 We hypothesized that more prolonged treatment with indomethacin may also decrease the occurrence of white matter injury on MRI.
The University of California San Francisco Committee on Human Research approved this prospective cohort study. Infants were studied after voluntary informed consent was obtained from parents for the MRI studies. Infants were eligible for this study if they were delivered before 28 weeks' gestation. Gestational age was based on the last menstrual period or early ultrasound; if the difference between the 2 methods was ≥7 days, the ultrasound date was used. Infants were excluded from the study if they had any of the following: (1) clinical evidence of a congenital malformation/syndrome; (2) congenital infections such as toxoplasmosis, rubella, cytomegalovirus, or herpes simplex infection; or (3) a large (>2-cm) parenchymal hemorrhagic infarction on ultrasound examination. Newborns with large parenchymal hemorrhagic infarction were excluded, because they are often too unstable to be safely transported to the MR scanner and frequently do not survive to hospital discharge.
Between October 1998 and October 2004, 216 premature newborns between 24 and 27 (+6 days) weeks' gestation at birth were admitted to our intensive care nursery, 177 were eligible for enrollment, and 57 (32%) participated in this study. The other 120 eligible infants were not enrolled because their parents were either unavailable or chose not to enroll their infant (n = 75), the infant was back transported to another hospital before the MRI could be obtained (n = 34), or the MRI magnet was not available for study use (n = 11). The 57 enrolled newborns were representative of all 216 admitted newborns of this gestational age, because they did not differ with respect to gestational age; birth weight; gender; and prenatal conditions, including preeclampsia, betamethasone treatment, gestational diabetes, and clinical chorioamnionitis, as well as neonatal conditions, including respiratory distress syndrome, chronic lung disease (defined as the need for supplemental oxygen at 36 weeks' PMA), necrotizing enterocolitis, symptomatic PDA, duration of indomethacin exposure, or need for surgical ductal ligation (all P > .1). Of the 57 newborns, 38 in this report were included previously in a cohort that examined the ability of white matter injury to predict early neurodevelopmental outcome.3 Our purpose in the current study was to identify the risk factors that are associated with white matter injury and moderate-severe brain abnormalities in a more consistently treated patient population with a smaller range of gestational ages and to examine the relationship of indomethacin with white matter injuries and brain abnormality (see below for definitions).
The newborns were studied with MRI as soon after birth as they were stable for transport to the magnetic resonance scanner (median age: 31.1 weeks' PMA, interquartile range: 29.9–32.6 weeks; median day of life: 37; interquartile range: 26–48 days). MRI studies were performed and scored using a protocol described previously.7 All of the newborns were monitored during scanning by a neonatologist in the magnetic resonance suite. The neonatologist hand ventilated intubated newborns and, when necessary, provided pharmacological sedation with pentobarbital and/or morphine sulfate according to our institution's sedation guidelines (29 newborns [51%]).
White matter injuries were considered to be foci of abnormal T1 hyperintensity in the absence of marked T2 hypointensity and foci of low intensity on T1-weighted images of the white matter. The severity of white matter injuries on MRI was classified as: normal (no white matter lesions), minimal (≤3 areas of T1 signal abnormality measuring <2 mm), moderate (>3 areas of T1 signal abnormality or these areas measured >2 mm but <5% of the hemisphere involved), or severe (>5% of the hemisphere involved).7,16 The interrater reliability of the grading system for white matter injuries was found previously to be high (κ = 0.84).7 Because severe white matter injury was seen in only 1 newborn, moderate and severe white matter injuries were grouped together. White matter injury scores were not assigned in 4 of the MRI scans because of excessive motion or other artifact.
In addition to the white matter injuries, we also examined the MRI scans for other abnormalities. We classified newborns as having moderate-severe “brain abnormality” on MRI if they had either moderate or severe white matter injuries or any degree of ventriculomegaly or severe IVH (grade 3 of Papile et al17). Mild ventriculomegaly was defined as the largest atrial ventricular diameter (at the level of the glomus of the choroid plexus) measuring 8 to 10 mm and moderate-severe as a ventricular diameter >10 mm.
Indomethacin Use in the Intensive Care Nursery
During the period covered by this study, all of the newborns delivered before 28 weeks' gestation were treated with a prophylactic course of indomethacin to prevent the development of a hemodynamically significant, symptomatic PDA. The prophylactic course of indomethacin was started within 15 hours of birth, and infants received 3 doses of indomethacin (0.2, 0.1, and 0.1 mg/kg) during a 48-hour period. A color Doppler evaluation was performed within 18 hours of the third dose of indomethacin to detect the presence of persistent flow through the ductus. If no flow was detected, the initial course was stopped after the third dose. If persistent flow was detected, then the initial course of indomethacin was extended by 3 additional doses (0.1 mg/kg at 24-hour intervals) unless there were contraindications to the additional use of indomethacin (serum creatinine >2.0 mg/kg or necrotizing enterocolitis). After the 3- to 6-dose initial indomethacin course, infants were monitored for the appearance of a hemodynamically significant, symptomatic PDA, indicated by the following 3 clinical symptoms: (1) prolonged systolic murmur, (2) bounding pulses, and (3) hyperdynamic precordium, in addition to echocardiographic evidence of left-to-right flow through the PDA. Nineteen of the 57 infants developed a hemodynamically significant, symptomatic PDA after the initial indomethacin course. All were treated with either a second course of indomethacin (3 doses) or surgical ligation. The decision to use a second course of indomethacin or to proceed directly to surgical ligation was made by the attending neonatologist. All of the newborns with a hemodynamically significant PDA ultimately required surgical ligation. We categorized the total exposure to indomethacin as short (1–3 doses) or prolonged (4–9 doses).15
Other Clinical Data
Other clinical data collected included variables associated previously with brain abnormalities in the premature newborn. A single neonatologist prospectively evaluated all of the newborns reported here and, together with a team of neonatal research nurses, prospectively collected data on the newborns' clinical condition during the admission to the intensive care nursery. Prenatal betamethasone included ≥2 doses 24 hours apart. Chorioamnionitis was diagnosed clinically (maternal fever >38°C during labor or fetal tachycardia with uterine tenderness, treated with antibiotics).18 Chronic lung disease was defined as a supplemental oxygen requirement after 36 weeks' adjusted gestational age. Postnatal infection was defined as culture-positive sepsis, pneumonia, or urinary tract infection. Hypotension was defined as a period of sustained low blood pressure requiring intervention with fluid bolus or inotropes.19–21 Necrotizing enterocolitis was diagnosed according to Bell's stage II criteria (presence of bloody stool with abdominal distension and persistent abnormal gastrointestinal radiology; evidence for intramural, portal venous, or free air; or evidence on surgical pathology).22
Clinical variables were compared with Stata 8 (Stata Corporation, College Station, TX) using the Mann-Whitney U test for continuous or ordinal data and the Fisher's exact test for categorical variables. Multivariate logistic regression with likelihood ratio testing was used to determine the independent association of variables that were significantly associated with moderate-severe white matter injuries on univariate analysis. We then tested the effect of these variables on the risk of white matter injury using a multivariate logistic regression with likelihood ratio testing adjusting for other clinical variables associated with white matter injury with a P < .3 on univariate analysis. These variables were also tested as independent predictors of moderate-severe brain abnormalities in a logistic regression model with likelihood ratio testing.
Moderate-severe white matter injury was detected in 12 (21%) of 53 newborns, minimal white matter injury in 13 (25%), and no white matter injury in 28 (53%). No newborn had cystic periventricular leukomalacia. On univariate analysis, prolonged indomethacin exposure and the presence of a hemodynamically significant PDA were the only 2 variables associated with moderate-severe white matter injury (Table 1). On univariate analysis, both of these risk factors were associated with a decreased incidence of moderate-severe white matter injury. None of the other perinatal or neonatal risk factors that we examined were significantly associated with moderate-severe white matter injury (Table 1).
All of the hemodynamically significant PDA had confirmed echocardiographic evidence of left-to-right flow through the PDA. As anticipated from prior reports,9 newborns with a hemodynamically significant PDA had higher rates of neonatal morbidity than infants without a symptomatic PDA, including necrotizing enterocolitis, chronic lung disease, and duration of mechanical ventilation (Table 2). In our nursery, most infants who develop a hemodynamically significant PDA also receive a second course of indomethacin.
Because the presence of a hemodynamically significant PDA was strongly associated with prolonged indomethacin exposure (P = .0001), we examined the effects of prolonged indomethacin exposure and presence of a hemodynamically significant PDA in a multivariate model to determine the independent effects of these 2 risk factors on white matter injury (Table 3). In this model, prolonged indomethacin exposure was still significantly associated with a reduced risk of moderate-severe white matter injury (odds ratio [OR]: 0.088; 95% CI: 0.002 to 0.99; P = .049). This corresponds with an 11-fold increase in the odds of white matter injury with a short course of indomethacin. However, the presence of a hemodynamically significant PDA was no longer associated with the risk of white matter injury (OR: 0.73; 95% CI: 0.02 to 22.2; P = .84). We then tested for an interaction between prolonged indomethacin exposure and hemodynamically significant PDA; the interaction term was not significant (P > .2).
Given the potential for confounding by other clinical variables, we examined the independent effects of prolonged indomethacin exposure and the presence of a hemodynamically significant PDA on the risk of white matter injury, adjusting for gestational age at birth, prenatal betamethasone, systemic infection, and days of mechanical ventilation. When adjusting for these other variables, prolonged indomethacin was even more strongly associated with a decrease in white matter injury (OR: 0.037; 95% CI: 0.004 to 0.69; P = .02). This corresponds with a 27-fold increase in the odds of white matter injury with a short course of indomethacin. In this expanded multivariate model, systemic infection also was significantly associated with the risk of white matter injury (OR: 6.1; 95% CI: 1.1 to 44.8; P = .04). The other variables included in the model were not significantly associated with the risk of white matter injury (all other variables: P > .5). Because chorioamnionitis and necrotizing enterocolitis were not seen in newborns with white matter injury, we were unable to include these variables in multivariate models.
Twenty (35%) of the 57 newborns met our definition of moderate-severe brain abnormalities: 11 newborns with moderate-severe white matter injury, 7 with ventriculomegaly, 1 with white matter injury and severe IVH, and 1 with ventriculomegaly and severe IVH. In a multivariate model including prolonged indomethacin exposure and the presence of a hemodynamically significant PDA, prolonged indomethacin exposure was significantly associated with a reduced risk of moderate-severe brain abnormalities (OR: 3.4 × 10−9; 95% CI: 0 to 0.29; P = .0002; Table 3). In contrast, a hemodynamically significant PDA was significantly associated with an elevated risk of moderate-severe brain abnormalities (OR: 9.7 × 107; 95% CI: 1.07 to ∞; P = .044). When adjusting for gestational age at birth, prenatal betamethasone, hemodynamically significant PDA, systemic infection, and days of mechanical ventilation, prolonged indomethacin exposure was similarly associated with a reduced risk of moderate-severe brain abnormalities (OR: 3.8 × 10−10; 95% CI: 0 to 0.15; P = .0009). In contrast, the other variables included in the model were not significantly associated with the risk of moderate-severe brain abnormalities (all P > .07).
In this observational study of newborns delivered before 28 weeks' gestation, the only neonatal factor that we found to be associated with a reduced risk of developing early MRI-detectable white matter injury was prolonged exposure to indomethacin treatment. Our study was limited by a relatively small sample size that, together with a large effect size, resulted in wide CIs on the likelihood ratio tests. Our study was also limited by a lack of randomization. Exposure to a prolonged course of indomethacin was determined by a newborn's PDA status rather than by randomization. In our study, newborns with a hemodynamically significant PDA had higher rates of neonatal morbidity than infants without a symptomatic PDA. The association between PDA status and neonatal morbidity has been demonstrated previously.9 Despite the lack of randomization and the likelihood of including more critically ill infants in the prolonged indomethacin exposure group, we found that prolonged indomethacin exposure was associated with a significant decrease in both moderate-severe white matter injury and brain abnormalities. In fact, when adjusting for the effect of other possible clinical factors, prolonged indomethacin exposure was even more strongly associated with a decrease in white matter injury. In the absence of randomization, however, we cannot exclude the possibility that unmeasured factors are confounding this association.
Indomethacin, a prostaglandin synthesis inhibitor, decreases cerebral blood flow, widens the range of cerebral vascular autoregulation, and matures the cerebral vascular basement membrane.9,23,24 Prior clinical trials have shown that a short (3 dose) course of prophylactic indomethacin reduces the occurrence of symptomatic PDA and severe IVH.9,10,12 Although individual trials have not demonstrated a reduction in white matter injury identified by ultrasound after a short course of indomethacin therapy,9 a recent meta-analysis of pooled results from 5 trials suggests that ischemic changes and periventricular leukomalacia may be decreased by prophylactic indomethacin.9 Despite these findings, short-course prophylactic indomethacin treatment has not been shown to improve neurodevelopmental outcomes.9,10,12
We have shown previously that a prolonged (6 dose) initial course of prophylactic indomethacin improves the rate of permanent ductus closure in infants who are at high risk for developing a hemodynamically significant symptomatic PDA.25 In a recent meta-analysis, a prolonged course of indomethacin treatment also was found to result in fewer cases of ductus reopening and a decreased rate of severe IVH.15 Our current findings suggest that prolonged exposure to indomethacin may also be associated with less moderate-severe white matter injury identified by MRI. In addition to the effects on cerebral hemodynamics, indomethacin also has antiinflammatory effects.23,26 White matter injury in premature newborns is associated with proinflammatory states, such as elevated proinflammatory cytokines in the cerebrospinal fluid,27 and requires several days before becoming apparent.1 We do not suggest that our current findings guide future clinical practice. Rather, at this time, we propose that a randomized trial of prolonged indomethacin treatment be conducted to determine whether indomethacin can interrupt the inflammatory process and decrease both white matter injury and subsequent neurodevelopmental abnormalities.
This research is supported by National Institutes of Health grants NS35902, NS40117, NS046432, HL466911, and HL56061 and by a gift from the Gates Foundation. Dr Miller is supported by the Canadian Institutes of Health Research Clinician Scientist Program (Phase 1). This study was conducted in part at the Pediatric Clinical Research Center, Moffitt Hospital, University of California San Francisco, with funds provided by the National Center for Research Resources, 5 M01 RR-01271, US Public Health Service.
We acknowledge the neonatal nurses of the Pediatric Clinical Research Center at University of California San Francisco (N. Newton, RN, MS; J. Bushnell, RN, MS; J. Ravitz Sturm, RN, BSN, MPH; J. Holland-Browne, RN; and J. Imamura-Ching, RN, BSN), whose skill and expertise made this study possible.
- Accepted October 17, 2005.
- Address correspondence to Steven P. Miller, MD, Department of Pediatrics, University of British Columbia, BC Children’s Hospital, 4480 Oak St, K3-180, Vancouver, BC, Canada V6H 3V4. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵Volpe JJ. Neurology of the Newborn. 4 ed. Philadelphia, PA: WB Saunders Company; 2001
- Maalouf EF, Duggan PJ, Counsell SJ, et al. Comparison of findings on cranial ultrasound and magnetic resonance imaging in preterm infants. Pediatrics.2001;107 :719– 727
- Inder TE, Anderson NJ, Spencer C, Wells S, Volpe JJ. White matter injury in the premature infant: a comparison between serial cranial sonographic and MR findings at term. AJNR Am J Neuroradiol.2003;24 :805– 809
- ↵Miller SP, Cozzio CC, Goldstein RB, et al. Comparing the diagnosis of white matter injury in premature newborns with serial MR imaging and transfontanel ultrasonography findings. AJNR Am J Neuroradiol.2003;24 :1661– 1669
- ↵Cooke L, Steer P, Woodgate P. Indomethacin for asymptomatic patent ductus arteriosus in preterm infants. Cochrane Database Syst Rev.2003;(2) :CD003745
- ↵Fowlie PW, Davis PG. Prophylactic intravenous indomethacin for preventing mortality and morbidity in preterm infants. Cochrane Database Syst Rev.2002;(3) :CD000174
- ↵Ment LR, Vohr B, Allan W, et al. Outcome of children in the indomethacin intraventricular hemorrhage prevention trial. Pediatrics.2000;105 :485– 491
- Ment LR, Vohr B, Oh W, et al. Neurodevelopmental outcome at 36 months' corrected age of preterm infants in the Multicenter Indomethacin Intraventricular Hemorrhage Prevention Trial. Pediatrics.1996;98 :714– 718
- ↵Fowlie PW, Davis PG. Prophylactic indomethacin for preterm infants: a systematic review and meta-analysis. Arch Dis Child Fetal Neonatal Ed.2003;88 :F464– F466
- ↵Herrera C, Holberton J, Davis P. Prolonged versus short course of indomethacin for the treatment of patent ductus arteriosus in preterm infants. Cochrane Database Syst Rev2004;(1) :CD003480
- ↵Versmold HT, Kitterman JA, Phibbs RH, Gregory GA, Tooley WH. Aortic blood pressure during the first 12 hours of life in infants with birth weight 610 to 4,220 grams. Pediatrics.1981;67 :607– 613
- ↵Ment LR, Stewart WB, Ardito TA, Huang E, Madri JA. Indomethacin promotes germinal matrix microvessel maturation in the newborn beagle pup. Stroke.1992;23 :1132– 1137
- ↵Quinn D, Cooper B, Clyman RI. Factors associated with permanent closure of the ductus arteriosus: a role for prolonged indomethacin therapy. Pediatrics2002;110 (1). Available at: www.pediatrics.org/cgi/content/full/110/1/e10
- Copyright © 2006 by the American Academy of Pediatrics