If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
The purpose of this study was to determine risk factors that are associated with hypoxic ischemic encephalopathy (HIE).
Study Design
This was a case-control study that included newborn infants with HIE who were admitted to the hospital between January 2001 and December 2008. Two control newborn infants were chosen for each case. Logistic regression and classification and regression tree (CART) analysis that compared control infants and cases with grade 1 HIE and control infants and cases with grades 2 and 3 HIE was performed.
Results
Two hundred thirty-seven cases (newborn infants with grade 1 encephalopathy, 155; newborn infants with grade 2 encephalopathy, 61; newborn infants with grade 3 encephalopathy, 21) and 489 control infants were included. Variables that were associated independently with HIE included higher grade meconium, growth restriction, large head circumference, oligohydramnios, male sex, fetal bradycardia, maternal pyrexia and increased uterine contractility. CART analysis ranked high-grade meconium, oligohydramnios, and the presence of obstetric complications as the most discriminating variables and defined distinct risk groups with HIE rates that ranged from 0–86%.
Conclusion
CART analysis provides information to help identify the time at which intervention in labor may be of benefit.
Intrapartum asphyxia in mature newborn infants causes 10-15% of cases of cerebral palsy, and its prevention is a major justification for the hospitalization of low-risk mothers who give birth in developed countries.
Despite advances in obstetric and neonatal care over the last 4 decades, the rate of cerebral palsy in normally formed newborn infants with a birthweight of >2.5 kg has not declined.
In addition, the seizure rate and the neonatal encephalopathy rate in newborn infants with a birthweight of >2.5 kg (which is often a marker of acute intrapartum neonatal brain injury) show no decline.
The objective of this study was to determine risk factors that are associated with the development of hypoxic ischemic encephalopathy (HIE).
Materials and Methods
Patient selection
Inclusion criteria were newborn infants who were born at the Rotunda Maternity Hospital in Dublin from January 2001 to December 2008 at ≥36 weeks 0 days' gestation and who required admission to the neonatal intensive care unit at ≤24 hours after delivery with evidence of encephalopathy. Newborn infants who were born between January 2001 and July 2005 were identified retrospectively. Newborn infants who were born between July 2005 and December 2008 were identified prospectively.
Grade of encephalopathy was assigned as the highest stage of encephalopathy (Sarnat and Sarnat
grading) that had been documented in the clinical notes and/or as noted on serial examination by a member of the research team (B.C.H., M.D.K., or S.M.). Two control newborn infants (the infants who were born before and after each case) were chosen for each case. Exclusion criteria for cases were out-born infants, <36 weeks' gestation, the presence of a major congenital anomaly, or any primary cause for encephalopathy other than hypoxia-ischemia. Exclusion criteria for control infants were out-born infants, <36 weeks' gestation, the presence of a major congenital anomaly, or any signs of encephalopathy in the neonatal period. If an infant was excluded as a control, then the infant who was delivered either before or after this infant was chosen.
The obstetrics definitions that were used in data acquisition are outlined in Table 1.
Table 1Obstetrics definitions used in data acquisition
Variable
Definition
Antenatal trauma
Significant fall, accident, or abdominal injury in the antenatal period
Late booking
Initiation of antenatal care at >24 weeks' gestation
Pregnancy-induced hypertension
Maternal blood pressure ≥140/90 mm Hg on 2 separate occasions >4 hours apart
Preeclampsia
New onset hypertension and proteinuria at >20 weeks' gestation
Proteinuria
>0.3 g protein/d in a 24-hour urine collection or, in the absence of a 24-hour urine collection, the presence of 2+ protein on dipstick
Gestational diabetes mellitus
Glucose intolerance with onset or first recognition during pregnancy and a normal glucose tolerance test by 6 weeks after delivery
Substantial antepartum hemorrhage
Vaginal blood loss equal to or greater than a menstrual period
Nonsubstantial antepartum hemorrhage
Vaginal blood loss less than a menstrual period
Fetal bradycardia
Decrease in the baseline fetal heart rate <100 beats/min
Late decelerations
Transient decrease in fetal heart rate that occurs at or after the peak of a uterine contraction
Fetal tachycardia
Increase in baseline fetal heart rate to ≥160 beats/min
Early decelerations
Transient decrease in fetal heart rate that coincides with the onset of a uterine contraction
Fetal heart rate variability
The beat-to-beat changes in fetal heart rate
Unsatisfactory cardiotocogram
The presence of a fetal bradycardia and/or late decelerations and/or fetal tachycardia and/or early decelerations (transient decrease in fetal heart rate that coincides with the onset of a uterine contraction) and/or fetal heart rate variability <5 beats/min
Satisfactory cardiotocogram
Baseline rate: 110-160 beats/min; moderate variability; absence of any late or variable decelerations; accelerations that may or may not be present
High-grade meconium
Grade 3 (thick or pea soup consistency) meconium or meconium that requires tracheal suction
Maternal pyrexia
Temperature ≥38°C measured with a tympanic thermometer
Duration of first stage of labor
The time from when the cervix was fully effaced and at least 1-cm dilated (in the presence of regular contractions) up to the time of full dilation
Shoulder dystocia
Difficult delivery of the shoulders that required additional obstetric maneuvers to release the shoulders after gentle downward traction failed
Uterine rupture
A defect that involves the entire uterine wall that was symptomatic and required surgical intervention
Placental abruption
Presence of retroplacental hematoma and clinical symptoms (as assessed by the clinical team at the time of delivery)
Hayes. HIE in newborn infants >36 weeks gestation. Am J Obstet Gynecol 2013.
Further details on data acquisition are available in the Appendix.
The data were analyzed by logistic regression analysis to identify the variables that were associated independently with HIE and with a classification and regression trees (CART) analysis to help define the distinct clinical groups at higher risk of HIE. CART analysis examines a dataset to find the best variables and associated cutoff points to group the data into those with and without the outcome in question. Splitting stops when the statistical process determines no further discriminating advantage with any of the remaining factors.
Two analyses were carried out: 1 analysis compared control infants and cases with grade 1 HIE; 1 analysis compared control infants and cases with grades 2 and 3 HIE. Criteria for inclusion included reaching statistical significance (P < .25) in the univariate analysis (Table 2, Table 3, Table 4) and clinical importance. Factors of clinical importance were defined as factors that have been associated with asphyxia and/or neonatal encephalopathy from previous published studies or deemed important from clinical practice. Logistic regression analysis was used to produce estimates of the odds ratios.
Table 2Univariate analysis of preconception, antenatal, peripartum, and neonatal factors
The first category of presentation looks at cephalic, breech and “other” (grouping all other types of presentation) as 3 separate groups. The second analysis groups breech with all other types of presentation
The first category of presentation looks at cephalic, breech and “other” (grouping all other types of presentation) as 3 separate groups. The second analysis groups breech with all other types of presentation
Occipitofrontal head circumference percentile for gestational age ≥97th percentile
No
446
95.3
145
96.7
49
83.1
17
94.4
< .05
Reference
Yes
22
4.7
5
3.3
10
16.95
1
5.6
1.55
0.79–3.00
Birth order
Singleton
486
99.6
153
97.5
60
98.4
21
100.0
.18
Reference
Twin
2
0.4
4
2.6
1
1.6
0
0
5.22
1.00–27.13
Sex
Female
241
49.4
75
47.8
22
36.1
9
42.9
.10
Reference
Male
247
50.6
82
52.2
39
63.9
12
57.1
1.24
0.90–1.69
36-37 wks' gestation
No
452
94.0
143
91.7
55
91.7
13
65.0
.001
Reference
Yes
29
6.0
13
8.3
5
8.3
7
35.0
1.86
1.06–3.25
ARM, artificial rupture of membranes; CI, confidence interval; NA, not available because of missing numbers; TENS, transcutaneous electrical nerve stimulation.
Hayes. HIE in newborn infants >36 weeks gestation. Am J Obstet Gynecol 2013.
a λ2 test for trend
b The first analysis looks at age ≤20 years and then in 5-year groupings up to ≥38 years. The second analysis simply looks at <25 years or >25 years
c χ2 test
d The first category of presentation looks at cephalic, breech and “other” (grouping all other types of presentation) as 3 separate groups. The second analysis groups breech with all other types of presentation
e The first analysis studies shoulder dystocia as a separate group, the second combines it with all other complications.
Ethical approval was obtained from the research ethics committee at The Rotunda Hospital.
Results
Two hundred forty-five cases and 490 control infants were included. Maternal records of 8 case newborn infants and 1 control newborn infant could not be located within the timeframe of data collection. Thus, results for 237 case newborn infants (155 newborn infants with grade 1 encephalopathy, 61 newborn infants with grade 2 encephalopathy, and 21 newborn infants with grade 3 encephalopathy) and 489 control newborn infants are described. An Apgar score of ≤5 at 10 minutes, a continued need for resuscitation (including endotracheal or mask ventilation) at 10 minutes after birth, and/or acidosis within 60 minutes of birth (defined as any occurrence of umbilical cord, arterial, or capillary pH ≤7.10) was present in 93 of 155 of newborn infants (60%) with grade 1 HIE; 47 of 61 newborn infants (77%) with grade 2 HIE, and 21 of 21 newborn infants (100%) with grade 3 HIE. Demographic characteristics of case newborn infants and control newborn infants are presented in Table 5.
Table 5Demographic characteristics of cases and control infants
Among the cases, 7 women (3%) did not have electronic fetal heart rate monitoring (2 women who underwent elective cesarean delivery and 5 women who proceeded to emergency cesarean delivery or spontaneous vaginal delivery without time for electronic fetal heart rate monitoring); intermittent electronic fetal heart rate monitoring was applied in a further 19 women (8%). All remaining cases had continuous fetal heart rate monitoring.
Among control infants, 50 women (10.2%) did not have electronic fetal heart rate monitoring (37 women who underwent elective cesarean delivery and 13 women who proceeded to emergency cesarean delivery or spontaneous vaginal delivery without time for electronic fetal heart rate monitoring); intermittent electronic fetal heart rate monitoring was applied in a further 84 women (17.2%). All remaining control infants had continuous fetal heart rate monitoring.
Results of univariate analysis are available (Table 2). The following factors reached statistical significance (P < .25) and were included in the multivariate analysis: maternal age, nationality, maternal smoking, history of infertility, primiparity, gestational age at booking, late booking, antenatal trauma, medication use in pregnancy, substantial antepartum hemorrhage, abnormal growth scans, artificial rupture of membranes, oligohydramnios, duration of first stage of labor, oxytocin use, maximum number of pains per 15 minutes, use of analgesia, general anesthesia, failure to reach full dilation, mode of delivery, maternal pyrexia, maternal antibiotics, obstetric complications (defined as the presence of shoulder dystocia, placental abruption, or uterine rupture), fetal heart rate abnormalities, high-grade meconium, birthweight, head circumference >97th percentile for gestational age, male sex, and gestation at delivery. Given that the decision to deliver by emergency cesarean delivery represents a number of factors, mode of delivery was not included in the CART analysis. Mode of delivery was included in logistic regression analysis.
Multivariate analysis
Comparison of control infants with cases of grade 1 HIE
Table 6 provides a summary of the logistic regression analysis. Seven factors (artificial rupture of membranes, oligohydramnios, >7 pains documented in any 15 minutes, maternal pyrexia, obstetric complications, high-grade meconium, and emergency cesarean and/or instrumental delivery) reached statistical significance for independent association with grade 1 HIE. Recognition of oligohydramnios and the presence of increased uterine contractility represent variables that may be recognized in early labor. Figure 1 shows the results from the CART analysis.
Table 6Logistic regression analysis of case control data
Included in the multivariate analysis to control for birthweight.
1.05
0.93–1.17
.449
0.98
0.82–1.18
.387
1.04
0.83–1.30
.702
0.82
0.603–1.10
.194
The following interactions were included in the multivariate analysis: primigravid × intervention at delivery; primigravid × bradycardia; stage 2 oxytocin × intervention at delivery; maternal pyrexia × intervention at delivery; maternal antibiotics × intervention at delivery; complications × intervention at delivery; complications × max number of pains per 15 minutes >7.
Analysis shows the hierarchy of factors, the percentage of HIE, and the number of records at each node. A, All subjects with grade 1 HIE and controls. B shows that the presence of an obstetric complication was the single most discriminating factor. The presence of C, high-grade meconium, D, oligohydramnios, and E, birthweight <2720 g helped to further discriminate between cases with grade 1 HIE and control infants.
CART, classification and regression tree; HIE, hypoxic ischemic encephalopathy.
Hayes. HIE in newborn infants >36 weeks gestation. Am J Obstet Gynecol 2013.
Comparison of control infants to grades 2 and 3 cases
Seven variables (significant antepartum hemorrhage, maternal pyrexia, fetal bradycardia, high-grade meconium, birthweight <3rd percentile for gestational age, head circumference >97th percentile for gestational age, and emergency cesarean/instrument delivery) reached statistical significance for independent association with grade 2 and 3 HIE on logistic regression analysis. Recognition of possible growth restriction and/or large fetal head size represents possible modifiable variables. Figure 2 shows the results from CART analysis. Some variables (maternal pyrexia, large head circumference) that have been found to be significant on logistic regression were not identified to form nodes with the tree approach. In CART analysis, the role of these factors may be buried within others because there is no attempt to identify independence, rather the goal is to define and rank the most predictive clinical groupings.
Figure 2CART analysis: comparison of control infants with cases with grade 2 and 3 disease
Analysis shows the hierarchy of factors, the percentage of hypoxic-ischemic encephalopathy, and the number of records at each node. A, All subjects with grade 2/3 HIE and controls. B shows that the presence of high-grade meconium was the single most discriminating factor. The presence of C, obstetric complications or D, oligohydramnios were the next most discriminating factors, followed by E, fetal bradycardia, F and J, male sex, G and I, birthweight, H, gestation <38 weeks, and K, maximum number of pains documented in any 15 minutes on partogram. Three terminal nodes show extreme rates of HIE. These are positions D, (high-grade meconium with oligohydramnios), F, (presence of obstetric complications and male sex), and K, (high-grade meconium, absence of oligohydramnios, birthweight >3590 g with >7 pains documented in any 15 minutes).
CART, classification and regression tree; HIE, hypoxic ischemic encephalopathy.
Hayes. HIE in newborn infants >36 weeks gestation. Am J Obstet Gynecol 2013.
This study identifies a number of important factors that place newborn infants at a significantly increased risk of not tolerating labor. These include higher grade meconium, growth restriction, a large head circumference, oligohydramnios, male sex, fetal bradycardia, and maternal pyrexia. A novel factor that was identified is increased uterine contractility. Defining terminal nodes and their associated complication rates demonstrates how CART analysis can provide useful information to help guide the time that intervention in labor may be of benefit.
The significant preconception/antenatal factors that were associated with encephalopathy that were found in this study differ from those in the Western Australian study.
This study did not find an association with a family history of seizures, maternal hypothyroidism, or neurologic problems. The broad definition of encephalopathy that was used in the study by Badawi et al
may have included newborn infants with encephalopathy of genetic or metabolic origin.
The lack of any association with socioeconomic status (ie, public/private booking, medical card status, and maternal employment) was surprising. This is in contrast to the findings of other studies.
In this study, the induction of labor by any method did not increase the odds of encephalopathy.
Frequency of contractions was a highly significant risk factor when there were >7 contractions during any 15-minute period. There are few studies on the influence of uterine activity on fetal status, despite awareness that increased uterine activity is associated with higher incidence of an umbilical artery pH of ≤7.11.
This study identified increased uterine contractions as an independent risk factor for asphyxia. In addition to frequency, the duration and amplitude of contractions are important
but were not available retrospectively. Uterine contraction rate over a 15-minute period was chosen because this information was easily available on the maternal partogram.
There is huge reliance on the cardiotocogram, despite poor specificity.
This study confirms the high false-positive rate with some abnormality noted on cardiotocogram in 68.9% of control newborn infants. In this study, fetal bradycardia was the only fetal heart rate abnormality that reached statistical significance for independent association with HIE. Although fetal heart rate trace is relied on heavily in the management of labor, uterine contractions (particularly if not associated with fetal heart rate changes) are often ignored. This study highlighted the limitations and usefulness of cardiotocogram and found that here is over reliance on fetal heart rate but less recognition of the importance of increased uterine activity.
In keeping with previous studies, maternal pyrexia in labor
were independent risk factors for the development of encephalopathy. It is suggested that hypoxia-ischemia and infection/inflammation share common inflammatory and molecular pathways and therefore have synergistic effects. Cerebral concentrations of proinflammatory cytokines have been shown to be elevated after exposure to either hypoxia-ischemia or infection/inflammation.
In this study, 57 of 724 women (7.9%) had pyrexia in labor, but positive cultures were identified in only 2 cases. Twenty of 57 women (35%) received antibiotics, which may have led to false-negative cultures, although these figures suggest that maternal pyrexia may be inflammatory rather than infective.
The association between fetal growth restriction and neonatal encephalopathy is well recognized.
In CART analysis, birthweight appeared in multiple branches, but with different thresholds. There are many different causes of fetal growth restriction, and each may differ in its potential to contribute to the development of encephalopathy.
A head circumference >97th percentile was associated independently with grade 2 and 3 HIE and remained so after adjustment for birthweight and gestation, which suggests that head size rather than overall size of the newborn infant may be the important factor. A previous study in rural Nepal found larger newborn infant head circumference carried a higher risk of neonatal asphyxia.
There was no independent association with higher newborn birthweight, which was similar to the findings of this study. A recent study examined the accuracy of sonographic estimation of fetal head circumference up to 3 days before delivery and noted that sonographic measurements consistently underestimated postnatal head circumference by an average of 13.5 mm or 4%.
This may be explained by the fact that sonographic measurements include only the bony calvaria of the fetal skull. In view of the high correlation between sonographic and postnatal measurements (r = 0.845; P < .001), fetal sonographic head circumference measurement before delivery may be of value.
Areas of interest for future studies include the impact of a large head circumference relative to maternal pelvic capacity and absolute head circumference measurement.
It is acknowledged that cause and effect are difficult to demonstrate in a case-control study. In addition, the retrospective component of this study was a limitation. Different datasets (intensive care admission records, ward journals, and radiology records) were cross-referenced, and it is unlikely that a significant number of newborn infants with encephalopathy during this period were missed. Labor and delivery data may be subject to bias because they were obtained from the maternal chart rather than prospectively. Data on maternal weight/pregnancy weight gain were not available, and labor information on the small number of women in the later stages of labor was limited. Fetal heart rate recordings on newborn infants with grades 2 and 3 encephalopathy (2001-2008) and with grade 1 encephalopathy (2004-2008) and 1 set of control recordings were reviewed independently by a blinded obstetrician (J.K.). Unfortunately, because a large proportion of fetal heart rate tracings were not located or had faded with time, only 102 fetal heart tracings were available. Because of the relatively small number of fetal heart rate recordings that were available, information on fetal heart rate tracings as documented in the medical chart at the time of delivery were used in the analyses. This may have limited the data that were available on fetal heart rate patterns.
All newborn infants with HIE were included regardless of demographic characteristics. Therefore, our results are applicable to any health care setting with similar demographics and health care systems.
The findings of this study may serve to identify infants who are at risk of not tolerating the labor process and assist in the decision-making process regarding the need for cesarean delivery.
Acknowledgments
We thank Myra O'Regan, Associate Professor of Statistics, Trinity College Dublin, for the statistical support and the children and their families who participated in this project.
Appendix.
Data acquisition
Preconception variables
Maternal age, nationality, family history, medical history (including thyroid and seizure history), obstetric history, parity, tobacco, alcohol, or medication exposure as recorded in the maternal chart was noted. Documented histories of infertility, in vitro fertilization, or clomiphene citrate use were used as markers of infertility. Parental occupations registered on booking and whether the mother held private health insurance, paid cash, or was publically funded were recorded as markers of socioeconomic status. With publically funded subjects, those who qualified for a medical card (awarded on the basis of income below a certain threshold or ongoing medical costs that would lead to undue hardship) were identified separately.
Antenatal variables
Apart from late bookers, all women had a dating ultrasound scan. When the menstrual cycle was regular and the last menstrual period was certain and crown-rump length was equivalent to the gestational age ±7 days, the expected date of delivery was calculated from the last menstrual period. In situations in which the menstrual cycle was irregular or the last menstrual period was uncertain or the crown-rump length differed from the gestational age at last menstrual period by >7 days, gestation was calculated by ultrasound scanning. The gestation at booking and any history of antenatal trauma in pregnancy were noted. Late booking and complications of pregnancy (in particular, pregnancy-induced hypertension, preeclampsia, gestational diabetes mellitus, or hypothyroidism) were recorded. Antepartum hemorrhage was graded as substantial or nonsubstantial, and the gestation at which it occurred was documented. Results of growth scans and biophysical profiles that had been performed antenatally were recorded.
Intrapartum and delivery variables
The intrapartum cardiotocogram features (if labor had been established) or the characteristics of the cardiotocogram that had been performed before delivery (if labor did not occur), as noted in the maternal chart by either the attending physician or midwife at the time of labor and/or delivery, were recorded. The amniotic fluid volume during the intrapartum period either on rupture of the membranes or by ultrasound assessment was recorded. The grade of meconium that had been noted at delivery was recorded; when this was not stated clearly, the level that had been noted before delivery was documented. The presence of maternal pyrexia and the use of maternal antibiotics in labor were noted. All analgesia that had been administered in the intrapartum period was documented. Use of spinal, epidural, and/or general anesthesia was recorded. The method of membrane rupture (spontaneous or artificial) and duration of rupture of membranes before delivery were noted.
The type of labor (spontaneous/induced/augmented) and the method used for induction or augmentation were recorded. The maximum rate of infusion of oxytocin at any period in the intrapartum period was noted; if the infusion had been discontinued, the reason for doing so was noted. The maximum number of contractions in any 15-minute interval on each partogram was documented.
For a subset of infants with grade 1 encephalopathy (those delivered in 2007 and 2008) and for all case infants with grades 2 and 3 encephalopathy, the total number of contractions as documented on the partogram was noted. For this variable, the total number of contractions as documented on the partogram for 1 set of control infants was recorded. The duration of first and second stages of labor (as recorded on maternal partogram) and whether a full 10 cm of dilation was reached was determined from maternal notes that included the maternal partogram. Patients with no labor were coded as length of first stage labor equal to zero. In addition, the mode of delivery for nonlaboring patients who had an elective cesarean delivery was coded with spontaneous vaginal delivery (0 = spontaneous vaginal delivery or elective section); those patients with an emergency cesarean delivery were coded with instrument deliveries (1 = forceps, ventouse, or emergency cesarean delivery). Details of any intervention during labor and delivery, the presentation of the infant in labor, details of complications that were encountered (shoulder dystocia, placental abruption, uterine rupture), and the mode of delivery were also recorded.
Infant factors
Infant gestation, birthweight, head circumference, Apgar score, cord pH, resuscitation details, and time to regular respirations were documented from the infant medical record. Neurologic signs on day 1 and 2 of life were recorded from the medical chart.
References
Himmelmann K.
Hagberg G.
Wiklund L.M.
Eek M.N.
Uvebrant P.
Dyskinetic cerebral palsy: a population-based study of children born between 1991 and 1998.
Funding for this study was provided by Friends of the Rotunda , an official fundraising arm and registered Charity (CHY240) of the Rotunda Hospital.
The authors report no conflict of interest.
Cite this article as: Hayes BC, McGarvey C, Mulvany S, et al. A case-control study of hypoxic-ischemic encephalopathy in newborn infants at >36 weeks gestation. Am J Obstet Gynecol 2013;209:29.e1-19.
00297-4&id=gr1.jpg){kind=link}
00297-4&id=gr2.jpg){kind=link}