Materials and Methods
Subjects
We conducted a case-control study of all neonates with suspected HIE who were admitted to our NICU for treatment with whole-body hypothermia from January 2007 to September 2012. This study was approved by our medical center Institutional Review Board. To qualify for whole-body hypothermia neonates had to be admitted and begin cooling within 6 hours of birth. Neonates were eligible for treatment with whole-body hypothermia if moderate-severe encephalopathy was present at birth that manifested as lethargy, stupor, coma, decreased or no activity, distal flexion, complete extension, decerebrate posture, hypotonia or flaccidity, abnormal primitive reflexes, bradycardia, periodic breathing, apnea, or seizures. Infants were eligible for whole-body hypothermia if they had a cord gas or early neonatal gas at <1 hour with pH ≤7.0, a base deficit >16 mM, cord/early neonatal gas at <1 hour with pH 7.01–7.15 and base deficit 10–15.9 mM (if moderate-severe encephalopathy was present with evidence of an acute sentinel event and a 10-minute Apgar score <5), or a need for assisted ventilation that was initiated at birth with continuation for at least 10 minutes. A sentinel event was defined as a signal event that occurred immediately before or during labor that could lead to fetal hypoxia such as ruptured uterus, abruption, massive fetomaternal hemorrhage, umbilical cord prolapse, ruptured vasa previa, amniotic fluid embolism, shoulder dystocia, or maternal cardiopulmonary arrest. Exclusion criteria included >6 hours of life, gestational age <35 weeks, severe growth restriction (birthweight, <1800 g), major congenital anomaly, severe persistent pulmonary hypertension with an anticipated need for extracorporeal membrane oxygenation, coagulopathy with active bleeding, and suspected sepsis with severe hemodynamic compromise that required large doses of pressors. Those neonates who were transported from outside institutions were started on passive cooling immediately on recognition of need for therapeutic hypothermia with instructions for temperature goal. All neonates were cooled to a rectal temperature of 33.5°C for 72 hours. After 72 hours of whole-body hypothermia, warming was initiated by increasing the set point of the automatic control on the servomechanism of the hypothermia system by 0.5°C per hour, which warmed the cooled blanket by 0.5°C per hour. All children who were enrolled in the hypothermia protocol were evaluated by a pediatric neurologist within 18 hours.
Infant and maternal medical records were reviewed to identify relevant clinical data. Intrauterine growth restriction was defined as an estimated fetal weight <10th percentile. Oligohydramnios was defined as an amniotic fluid index <5.0 cm with intact membranes at the time of the admission in which the delivery occurred. Preeclampsia was defined as proteinuria, edema, and the presence of new onset hypertension. The clinical diagnosis of chorioamnionitis was made in the presence of maternal fever, with at least 1 other finding of fetal tachycardia, uterine tenderness, or purulent vaginal discharge. Patients who were diagnosed with clinical chorioamnionitis were started immediately on intravenous antibiotics. Sepsis was considered present only for neonates with positive blood and/or cerebrospinal fluid cultures. The diagnosis of nonreassuring fetal heart rate tracing was made by the physician who attended the delivery before performing a vacuum, forceps, or cesarean delivery.
Imaging
A brain MRI with diffusion tensor images was performed between day of life 7 and 10. Cases had severe brain injury, defined as an abnormal brain MRI or death. Control subjects were surviving neonates with normal MRI. These MRIs were reviewed by an experienced pediatric neuroradiologist at our institution (T.A.G.M.H.). The images were reviewed for focal or diffuse lesions related to hypoxic-ischemic injury. Neuroimaging abnormalities of an acute perinatal insult were defined as brain swelling; cortical highlighting; focal or global loss of grey-white matter differentiation; abnormal signal intensity in the basal ganglia and thalami; loss of normal signal intensity in the posterior limb of the internal capsule; acute and subacute parenchymal, intraventricular, or extracerebral hemorrhage; and acutely evolving focal infarction in an arterial territory or in a parasagittal or watershed distribution.
Statistics
Univariate analysis was performed with the Student t test to evaluate continuous variables, and χ 2 or Fisher exact test for categoric variables. A probability value of < .05 was considered statistically significant. Logistic regression models were created with forward stepwise selection of risk variables with a probability value of < .1 and biologic plausibility after elimination of similar confounding variables. Models were used to determine the predictive value of factors that were hypothesized to impact severe brain injury. Receiver operating characteristic curves were generated, and sensitivity and specificity were calculated. Stata software was used for statistical analysis (version 10; StataCorp LP, College Station, TX). Cross-validation analysis was performed by dividing the participants evenly into 5 groups, combining 80% of the participants, or 4 groups at a time, and creating a model based on those 4 groups. Logistic regression was then run on the aggregate of 4 groups and used to predict the outcome for the fifth group. This process was repeated until each group had been the test group.
Results
During this 5-year 8-month period, there were 109 neonates admitted to our NICU with suspected HIE for treatment with whole-body hypothermia; 39% were born at 1 of 2 hospitals within our system, and 61% were transferred from outside institutions within our state. Of these 109 neonates, 98 (90%) had an MRI performed at 7-10 days of life; 8 died without having imaging; 1 had a computed tomography scan, and 2 survivors were discharged before having an MRI performed. Fifteen neonates were missing data on the commencement of spontaneous respirations, and 2 neonates were missing Apgar scores. The 2 patients for whom we are missing Apgar scores were both born unintentionally at home. Cord pH at birth was available for 84.7% of cases and 85.2% of control subjects; cord base deficit was available for 71.7% of cases and 82.0% of control subjects. Cord gases were not always done routinely at birth in referring institutions. Of the 8 neonates who died without imaging, 6 were emergent cesarean deliveries for nonreassuring fetal heart rate tracing; 3 had sentinel events that included abruption, uterine rupture, and maternal cardiac arrest, and 5 had seizures. The 8 neonates who died before having an MRI were diagnosed with Sarnat stage III encephalopathy before starting cooling within 6 hours of life. The 46 cases with an abnormal brain MRI or death were compared with 60 survivors with a normal MRI. In the case group, cerebral white matter injury occurred in 29 of 46 neonates (63%), and basal ganglia/thalamic injury occurred in 21 neonates (45.7%). Areas of infarcted white matter were noted in 6 neonates (13%). Extracerebral hemorrhage occurred in 10 neonates (21.7%): 9 had subdural hematoma, and 1 had cephalohematoma, all of which were associated with either cerebral white matter and/or basal ganglia injury. Acute and subacute parenchymal and intraventricular hemorrhage occurred in 7 neonates (15.2%), always in conjunction with other abnormalities.
In univariate analysis, cases and control subjects did not differ on gestational age, sex, race, birthweight, mode of delivery, sentinel events, positive neonatal blood cultures, or diagnosis of nonreassuring fetal heart rate tracing before delivery ( Tables 1 and 2 ). There were 33 cesarean deliveries performed in the case group, of which 29 deliveries were done for nonreassuring fetal heart rate tracing, 1 emergency cesarean delivery was done after a motor vehicle accident, 1 emergency cesarean delivery was done after a maternal seizure, 1 cesarean delivery was done after a failed vacuum delivery of a 4140-g infant, and 1 cesarean delivery was done for a previous myomectomy at which time there was difficulty extracting the infant for 10 minutes, which required vacuum assistance and the extension of the uterine incision. There were 35 cesarean deliveries in the control group, of which 30 were done for nonreassuring fetal heart rate tracing, 1 was an elective repeat cesarean during which the placenta was incised during delivery, 1 for arrest with chorioamnionitis and fetal tachycardia, and 3 for arrest with unremarkable fetal heart rate tracings. There was 1 forceps delivery in the case group and 3 in the control subjects, all with nonreassuring fetal heart rate tracing noted before delivery. There were 2 vacuum deliveries in the case group, both with nonreassuring fetal heart rate tracing; there were 6 vacuum deliveries in the control subjects of which 3 infants had nonreassuring fetal heart rate tracing. Of those infants with meconium-stained fluid noted at delivery, 4 of 17 cases (24%) and 9 of 25 control subjects (36%) were noted to have thick meconium ( P = .39). Cases were significantly more likely to have had an abruption ( P = .03), 5-minute Apgar score <5 ( P = .049), higher initial white blood cell count ( P = .025), lower initial neonatal platelet counts ( P = .03), require chest compressions ( P = .02) and intubation ( P = .02) in the delivery room, first spontaneous respiration at >30 minutes of life ( P = .001), and seizures ( P = .03; Tables 1 and 2 ). Cases were significantly more likely to have metabolic acidosis with pH <7.0 and a base deficit of >12 mM on the cord gas at birth or the initial neonatal arterial gas. The incidence of the initial neonatal arterial pH being <7.0 was 20 of 46 for the cases (43.5%) and 8 of 60 for the control subjects (13.3%; P = .0005).
| Variable | Cases (n = 46) | Control subjects (n = 60) | P value |
|---|---|---|---|
| Maternal age, y | 29.1 ± 8.2 | 26.5 ± 6.8 | .08 |
| Nulliparous, n (%) | 27 (59) | 34 (57) | .80 |
| Race, n (%) | .80 | ||
| White | 19 (41) | 25 (42) | |
| Black | 20 (43) | 29 (48) | |
| Hispanic | 3 (7) | 3 (5) | |
| Other | 4 (9) | 3 (5) | |
| Cesarean delivery, n (%) | 33 (72) | 35 (58) | .45 |
| Multiple birth, n (%) | 0 | 1 (2) | 1.00 |
| Preterm premature rupture of membranes, n (%) | 0 | 1 (2) | 1.00 |
| Preeclampsia, n (%) | 3 (7) | 7 (12) | .51 |
| Antenatal magnesium, n (%) | 2 (4) | 2 (3) | 1.00 |
| Tobacco use, n (%) | 2 (4) | 1 (2) | .58 |
| Cocaine use, n (%) | 2 (4) | 4 (7) | .70 |
| Intrauterine growth restriction, n (%) | 3 (7) | 5 (8) | 1.00 |
| Oligohydramnios, n (%) | 2 (4) | 3 (5) | 1.00 |
| Abruption, n (%) | 12 (26) | 6 (10) | .03 a |
| Meconium, n (%) | 17 (37) | 25 (42) | .69 |
| Oxytocin for labor induction/augmentation, n (%) | 8 (17) | 21 (35) | .05 |
| Nonreassuring fetal heart rate tracing, n (%) | 32 (70) | 37/58 (64) | .40 |
| Sentinel event, n (%) | 20 (43) | 23 (38) | .59 |
| Clinical chorioamnionitis, n (%) | 5 (11) | 9 (15) | .53 |
| Variable | Cases (n = 46) | Control subjects (n = 60) | P value |
|---|---|---|---|
| Gestational age, wk | 38.6 ± 1.5 | 39.0 ± 1.7 | .17 |
| Birthweight, g | 3224 ± 659 | 3256 ± 597 | .79 |
| Female sex, n (%) | 18 (39) | 22 (37) | .80 |
| 5-minute Apgar score <5, n (%) | 34 (74) | 32 (55; n = 58) | .049 a |
| Cord pH | 6.85 ± 0.15 (n = 39) | 6.98 ± 0.14 (n = 52) | < .0001 a |
| Cord base deficit, mM | 19.6 ± 6.6 (n = 33) | 13.5 ± 7.5 (n = 50) | .0003 a |
| Cord pH <7.0 and base deficit >12 mM, n (%) | 36 (88; n = 41) | 37 (70; n = 53) | .038 a |
| First arterial pH | 7.05 ± 0.20 | 7.17 ± 0.15 | .0009 a |
| First arterial base deficit, mM | 20.5 ± 9.5 (n = 45) | 15.4 ± 7.9 | .0035 a |
| Neonatal length of stay, d | 21.3 ± 15.8 | 15.9 ± 9.2 | .03 a |
| Respiratory distress, n (%) | 13 (28) | 24 (40) | .21 |
| Intraventricular hemorrhage, n (%) | 2 (4) | 1 (2) | .58 |
| Hyperbilirubinemia, n (%) | 4 (9) | 12 (20) | .17 |
| Seizures, n (%) | 25 (54) | 20 (33) | .03 a |
| Initial neonatal white blood cell count, n | 20,694 ± 8493 | 17,190 ± 7301 | .025 a |
| Initial neonatal platelets, K/cu mm | 158 ± 49 | 181 ± 58 | .03 a |
| Initial neonatal hematocrit, % | 47.2 ± 8.1 | 44.4 ± 8.1 | .08 |
| Glucose level, mg/dL | 124 ± 130 | 108 ± 63 | .43 |
| Days to oral feeding | 29.1 ± 61.6 (n = 29) | 12.9 ± 20.1 (n = 59) | .07 |
| Chest compressions in delivery room, n (%) | 23 (50) | 17 (28) | .02 a |
| First respiration at >30 minutes, n (%) | 29 (74; n = 39) | 21 (40; n = 52) | .001 |
| Intubation in delivery room, n (%) | 40 (87) | 40 (67) | .02 a |
| Positive blood culture, n (%) | 3 (6) | 1 (2) | .31 |
Additional variables with a probability value of <.1 after univariate analysis included maternal age ( P = .08), absence of intrapartum exposure to oxytocin ( P = .05), and neonatal length of stay ( P = .03). Cord pH, cord base deficit, and initial neonatal arterial base deficit were not included in the multivariable regression model because of confounding with initial neonatal arterial pH. Initial neonatal arterial blood gas was chosen over cord blood gas because more umbilical cord data were missing. Intubation in the delivery room was not included because it is performed routinely when chest compressions are administered and may impact spontaneous respirations. Given that our goal was to examine early predictors of neurologic injury (within 1 hour of life), neonatal length of stay was not included in the model. Therefore, the following variables were included in the multivariable analysis: maternal age; oxytocin exposure in labor; abruption; 5-minute Apgar score <5; chest compressions in delivery room; initial neonatal arterial pH; initial neonatal white blood cell count, platelets, and hematocrit level; neonatal seizures, and spontaneous respirations at >30 minutes.
In multivariable logistic regression, initial neonatal arterial pH ( P = .004), spontaneous respiration at >30 minutes of life ( P = .002), and absence of intrapartum exposure to oxytocin ( P = .033) predicted a significantly increased risk of abnormal brain MRI at 7-10 days or death ( Tables 3 and 4 ). All patients had contractions, and the risk of abruption was significantly lower for patients who received intrapartum oxytocin (1/30 women; 3.3%), compared with those without oxytocin exposure (17/61 women; 27.9%; P = .02). For those women without intrapartum exposure to oxytocin, the cesarean delivery rate was 50 of 77 women (64.9%); for those women who were exposed to oxytocin, the cesarean delivery rate was 18 of 29 women (62.1%), which was not significantly different ( P = .78) and showed that the threshold for performing a cesarean delivery seems to be roughly equivalent in both groups.
| Variable | Coefficient | 95% CI | P value |
|---|---|---|---|
| Maternal age | 0.01 | –0.003 to 0.023 | .13 |
| Intrapartum oxytocin exposure | –0.23 | –0.44 to –0.01 | .04 a |
| Abruption | 0.07 | –0.014 to 0.157 | .16 |
| 5-minute Apgar score <5 | –0.0001 | –0.24 to 0.24 | 1.00 |
| Chest compressions in delivery room | –0.06 | –0.31 to 0.19 | .65 |
| Initial neonatal arterial pH | –0.73 | –1.28 to –0.18 | .01 a |
| Initial neonatal white blood cells | 8.78e–06 | –3.61e–06, 2.12e–05 | .16 |
| Initial neonatal hematocrit level | 0.007 | –0.005 to 0.019 | .24 |
| Initial neonatal platelets | –0.002 | –0.0033 to 0.0002 | .08 |
| Neonatal seizures | 0.08 | –0.12 to 0.28 | .43 |
| Spontaneous respirations at >30 min | 0.27 | 0.05–0.50 | .02 a |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
