Umbilical artery blood gas parameters in neonates with early onset seizures who die


Dr K. Williams, Department of Obstetrics and Gynecology, Yale University School of Medicine, 333 Cedar Street, P.O. Box 208063, New Haven, Connecticut 06520-8063, USA.


Objective  To relate umbilical artery blood gas parameters to mortality among neonates with hypoxic–ischaemic encephalopathy related to early onset seizures.

Design  Population cohort study.

Setting  British Columbia Women's Hospital.

Population  Forty-seven infants at ≥32 weeks of gestation admitted to NICU with early onset seizures secondary to hypoxic–ischaemic encephalopathy with umbilical artery blood gases done at delivery.

Methods  Patients were divided into two groups: (1) Infants with neonatal seizures who survived, and (2) infants with neonatal seizures who died related to hypoxic–ischaemic encephalopathy complications. Comparison of umbilical artery pH, PO2, PCO2, base deficit was done between the two groups with Student's t tests.

Main outcome measures  Umbilical artery pH, PO2, PCO2 and base deficit.

Results  The PO2 was significantly higher in the group that expired (18.36 ± 9.15 vs 12.33 ± 7.51). There were no significant differences in any other blood gas parameters between the groups.

Conclusion  Neither the umbilical artery pH nor base deficit is predictive of neonatal death in infants with hypoxic–ischaemic encephalopathy with seizures. The finding of a high PO2 in neonates who died may indicate an inability of those infants to efficiently extract oxygen from blood.


Hypoxic–ischaemic encephalopathy is diagnosed in neonates with the abnormal neurological findings such as jitteriness, seizures, hypo/hypertonia and coma in the setting of severe asphyxia.1–3 The American College of Obstetricians and Gynecologists4 define hypoxia severe enough to cause hypoxic–ischaemic encephalopathy as a constellation of findings including pH <7.0, Apgar scores of 0–3 for longer than 5 minutes, in association with neonatal neurologic sequelae (e.g. seizures, coma, hypotonia) and any of the following, cardiovascular, gastrointestinal, haematologic, pulmonary or renal system dysfunction. Although various scoring systems are used.2,3 early onset seizures is the most clearly defined parameter to indicate moderate to severe disease with hypoxic–ischaemic encephalopathy.5–11 Neonates with hypoxic–ischaemic encephalopathy have an increased risk of long term problems, including cerebral palsy and neonatal death.5,6,12,13 The parameters of Apgar score, umbilical artery pH and base deficit levels are used by investigators to predict hypoxic–ischaemic encephalopathy and subsequent long term neurological outcome.5–12 Whether these parameters are also predictive of neonatal death in patients with documented moderate to severe hypoxic–ischaemic encephalopathy is unknown. We therefore correlated umbilical artery blood gas parameters with the risk of neonatal death secondary to hypoxic–ischaemic encephalopathy.


Institutional approval for this study was granted by BC Women's Hospital research review board. We reviewed the maternal and neonatal charts of 170 patients admitted to the special care nursery (Neonatal Intensive Care Unit) at BC Women's Hospital in Vancouver, British Columbia over a 10-year period from January 1990 to December 1999 with a neonatal diagnosis of hypoxic–ischaemic encephalopathy. The inclusion criteria for our study were:

  • Gestational age ≥32 weeks at delivery.

  • Patients having a documented seizure in the first 24 to 48 hours after admission to the NICU secondary to hypoxic–ischaemic encephalopathy. The neonatal seizures were evaluated to ensure that they were unrelated to trauma, febrile illness or intracranial mass. In our centre, hypoxic–ischaemic encephalopathy is diagnosed in babies with asphyxia using the guidelines of Sarnat and Sarnat.2 Babies with hypoxic–ischaemic encephalopathy underwent EEG assessment to document the presence or absence of electrographic epileptiform discharges, disturbed background activity, burst suppression or extremely low voltage patterns. Ancillary documentation that the seizure was related to hypoxic–ischaemic encephalopathy was sought in each case by reviewing any available CT findings.

  • The patients had umbilical artery blood gas analysis done at delivery. The primary neonatal outcome was the occurrence of death related to complications associated with hypoxic–ischaemic encephalopathy. Our review of the literature revealed that Andres et al.14 identified that the mean base deficit in neonates with seizures was (20.6 ± 4.93) with an increase in a base deficit to 24 in patients with hypoxic–ischaemic encephalopathy. We conducted a power analysis to identify an increase of five units in base deficit in patients with seizures who survived compared with patients with seizures who expired. The intended sample size was based on a 1:2 ratio of infants who died compared with those who survived with an expected 80% power and a type 1 error of 5%. We anticipated we would need 15 patients in the expired group versus 30 patients in the group that survived.

Comparison of the umbilical artery blood gas continuous variables of pH, Po2, Pco2 and base deficit was done between those neonates that died and those that survived using unpaired Student's t tests. A P value of less than 0.05 was considered to be statistically significant.


One hundred and seventy neonatal charts with a neonatal diagnosis of hypoxic–ischaemic encephalopathy were reviewed and 47 patients fit the entry criteria. Of these, 14 neonates died of complications related to hypoxic–ischaemic encephalopathy. From the 170 neonatal charts, we excluded 47 patients because they were out born, 25 patients because they were <32 weeks, 20 patients because no cord gases were done at delivery because of physician error and 31 patients because although they had a neonatal diagnosis of hypoxic–ischaemic encephalopathy, they did not have seizures. Active resuscitation after birth was attempted on all neonates included in the study. There was no significant difference between the maternal demographics of the groups (Table 1).

Table 1.  Maternal demographics. Values are presented as mean (SD) or percentage (%).
Maternal age (years)33 (4)34 (5)NS
Gravidity3 (1.2)2.3 (1.4)NS
Gender (male)5346NS
Birthweight (g)3379 (532)3426 (553)NS
Delivery (caesarean section)7167NS
Gestational age (weeks)39 (1.3)40 (1.2)NS

The umbilical gas parameters of neonates who died were compared with those that survived (Table 2). There were no significant differences in the umbilical artery blood gas parameters of pH, base deficit, Pco2 and O2 Sat between the groups.

Table 2.  Correlation of umbilical artery parameters of babies who died and those who survived. Values are presented as n or mean (SD).
Cord pH6.89 (0.15)6.94 (0.17)NS
Base deficit (mmol/L)−17.04 (5.34)−13.95 (9.72)NS
PCO2 (mmHg)109 (31.88)94.9 (36.27)NS
PO2 (mmHg)18.36 (9.15)12.33 (7.51)0.04
O2 Saturation0.16 (1.9)1.42 (6.65)NS

The umbilical artery cord Po2 levels were significantly higher in the group that died compared with the group that survived (18.36 ± 9.15 vs 12.33 ± 7.51) (P < 0.04) (Table 2). Figure 1 shows the distribution of base deficit between the two groups.

Figure 1.

Box plots of base excess in a sample of neonates by outcome. Box plots demonstrate median 25th and 75th centiles and smallest and largest values.


Low et al.15 determined that the threshold for fetal metabolic acidosis at delivery, which predicts newborn complications, is an umbilical artery base deficit of −12 to −16.

Other literature reports6,12,13,16,17 indicate that neurological dysfunction and other complications correlate with an umbilical artery pH of less than 7.0 and a 5-minute Apgar score less than 3. Goodwin et al.13 documented an incidence of encephalopathy of 31% (with an umbilical artery pH < 7). Neonatal seizures occurred in 9% of infants with an umbilical artery pH between 6.9 and 6.99 and in 80% of infants with an umbilical artery pH between 6.61 and 6.70. There were three stillbirths and five neonatal deaths in this population.

Andres et al.14 used umbilical pH to review severe neonatal morbidity and death. He identified that neonates who died had an increased base deficit and elevation of their bicarbonate levels but only three neonates died secondary to hypoxic–ischaemic encephalopathy and so he could not provide further analysis to help elucidate this problem.

Belai et al.18 identified umbilical artery and venous Pco2 differences of a Pco2≥ 25 Torr, as a highly sensitive (89%) and predictive (56%) parameter in identifying the risk of neonatal death in patients with abnormal neurological function immediately post-delivery (i.e. seizures).

We reviewed 170 neonates reported as having hypoxic–ischaemic encephalopathy at a provincial tertiary referral hospital with approximately 7500–8000 deliveries a year. We included only patients at a gestational age of ≥32 weeks because in our nursery the neonatal mortality in this group approaches that of the term infant, and in some of infants delivered prior to 32 weeks with suspected hypoxic–ischaemic encephalopathy, the investigator could not appropriately identify whether the neonatal death occurred secondary to the effects of prematurity or hypoxic–ischaemic encephalopathy. In our centre, 30 of our inborn cases did not have cord gases done at delivery and so were not included. In addition, all out-born cases were excluded because of concerns about other factors including initial stabilisation and neonatal transport, which may affect the outcome.

Because of the significant variation in the literature as regards to appropriate diagnosis of hypoxic–ischaemic encephalopathy, we chose to use neonatal seizures which is uniformly present in patients with a diagnosis of moderate hypoxic–ischaemic encephalopathy. With the use of seizures, we hope to define a significant high risk neonatal problem. We were careful to exclude all patients in which there was another cause for the seizures.

Even though our study was conducted over a 10-year period, we identified only 14 neonates who died secondary to hypoxic–ischaemic encephalopathy to include in our study. Our study is the largest we could find to correlate neonatal death with umbilical artery parameters.

One unexpected finding in this study is that umbilical artery Po2 was elevated in the group of infants that died. A previous report had indicated that Po2 is of no value in the prediction of convulsions.13 One possible explanation for the elevated Po2 is highlighted by Ward17 who identified a previously unrecognised threat to tissue oxygenation relevant in the presence of hypercapnia acidosis. Hypercapnia handicaps cell's capacity to maintain energy status anaerobically and enhances the risk of hypoxic injury during anaemia or other situations in which oxygen delivery is critically reduced. Significant impairment of oxygen extraction occurred during severe hypercapnia. Impairment of O2 extraction was worse at Pco2≥ 118. We did not encounter any neonates with a pH of less than 6.60, as this level of acidosis may be incompatible with fetal survival.

Our study showed that neither umbilical artery pH levels nor base deficit will predict which neonates will die.

A search for more sensitive predictors must continue. Engle et al.19 reviewed infants with marked acidaemia at birth who developed convulsions and were ventilated. Their findings that these neonates had a higher Pco2 than the normal group with a significant decrease in Pco2 within 2 hours requires further investigation. Patterns of change of Pco2 may be predictive of the development of neonatal encephalopathy. Changes in acute blood gas parameters in the hours after birth need to be explored in future studies to predict short term and long term neonatal outcome in asphyxiated infants.

Accepted 26 March 2004