While most fetuses experience the physiologic fetal-to-neonatal transition to air breathing after birth at term, 10% of newborns require some degree of active resuscitation to activate breathing and 1% require extensive care [1]. According to the most recent neonatal resuscitation protocol endorsed by the American Academy of Pediatrics (AAP) and supported by published studies, it was reported that approximately 3% of newborns needed positive-pressure ventilation, 2% needed endotracheal intubation, and 0.1% needed cardiac compressions or epinephrine administration [234].

Difficulty in initiating respiration, and depression of tone and reflexes are common symptoms of neonatal encephalopathy (NE), which is a clinically defined syndrome of disturbed neurologic function in the earliest days of life in babies born at or beyond 35 weeks of gestation manifested by a subnormal level of consciousness or seizures. Historically, the incorrect assumption that most NE results from hypoxia during the intrapartum period has hindered serious research into other possible causes of NE. Recent epidemiologic studies have identified several preconception demographic, and maternal medical conditions (advanced maternal age, family history of seizure, family history of neurologic disorder, maternal thyroid disease, and infertility treatment) and antepartum risk factors (preeclampsia, moderate to severe vaginal bleeding, advancing pregnancy beyond 39 weeks gestation, late or no prenatal care, and low neonatal birth weight percentile) as independent risk factors for NE [567]. Of note, it was reported that 70% of NE cases were likely the result of events arising before the onset of labor [8].

Hypoxic ischemic encephalopathy (HIE) comprises a cause-specific subset of all NE. Inaccurate prediction of fetal acidemia by intrapartum fetal heart rate (FHR) monitoring can lead to the presumptive diagnosis of HIE as the direct cause of depressed neonates in cases that are otherwise unexplained, thereby overestimating the true incidence of HIE. According to the American College of Obstetricians and Gynecologists (ACOG) and the AAP, a low cord arterial blood pH <7.0 and/or a base excess (BE) <−12 mmol/L are the generally accepted cutoff values for pathological acidosis, which increases the risk of seizures, HIE, and cerebral palsy (CP) [9]. As umbilical cord blood gas and acid-base assessment are the most objective predictors of fetal acidemia [1011], it is usually recommended by the ACOG to obtain a cord arterial blood gas analysis (ABGA) in certain clinical circumstances such as a low 5-minute Apgar score, severe growth restriction, or abnormal FHR tracing [1213], although there is no specific guideline in Korea.

Meanwhile, as a tertiary center, our institution has adopted the routine performance of cord blood ABGA at birth for every delivery since 2006 for a variety of reasons, including medicolegal considerations. Surprisingly, we could not find any study with a large volume of data including Korean population regarding cord ABGA at birth and the association analysis between cord pH and other clinical factors such as meconium staining, nuchal cord, and mode of delivery. This led us to initiate a comprehensive analysis of cord blood pH in consecutive singleton pregnancies delivered at term in our hospital population.

The specific questions that we pursued here were as follows. First, what are the distribution of cord blood pH and the incidence of low cord pH? Second, what are the associations between low cord blood pH and clinical factors including the grade of meconium, number of nuchal cord cases, and mode of delivery? Thirdly, among cases with low cord pH defined as a pH <7.1, what is the actual frequency of adverse neonatal outcomes including neonatal intensive care unit (NICU) admission, ventilator support, need for brain ultrasonography, and abnormal sonographic findings? Lastly, how many CP cases were subsequently diagnosed among neonates with low cord pH at term?

During the study period from January 2006 to December 2016, there were 21,428 deliveries in our institution including 4,855 preterm deliveries, 16,554 term deliveries, and 19 post-term deliveries. Fig. 1 shows the inclusion and exclusion protocols of this study population. We excluded cases with preterm birth, post-term birth, twin pregnancy including vanishing twin, stillbirth and fetal death in utero. After the exclusions, there were 15,701 term singleton pregnancies. Among them, the results of umbilical cord ABGA were available in 14,221 cases (90.6%), which constituted the final population of this study. We collected data on clinical variables including maternal age, height, weight, parity, history of prior preterm delivery, gestational age at delivery, the presence of meconium staining, the number of nuchal cord, mode of delivery, sex of baby, cord blood pH, Apgar scores at 1 minute and 5 minutes, neonatal birth weight, and NICU admission. Data were acquired from the electronic medical record system in our institution. The grade of meconium staining was divided into the following 5 groups by experienced labor and delivery nurses: none, 1+, 2+, 3+, and 4+; nuchal cord was graded on a scale of 0 to 4. Mode of delivery was divided into 4 groups: elective caesarean section delivery, emergent caesarean section delivery, spontaneous vaginal delivery, and vacuum-assisted vaginal delivery. Apgar scores were assigned by experienced labor and delivery nurses or pediatric doctors. The cord blood was sampled immediately after delivery with heparinized syringes and analyzed within 30 minutes in most cases. We defined low cord pH as a pH <7.1. Neonatal weight was categorized into 3 groups: small for gestational age (SGA), appropriate for gestational age (AGA), and large for gestational age (LGA) according to national data from the Korean Health Insurance Review and Assessment Service 2009.

In cases of low cord pH, we also examined additional neonatal and infant outcomes including ventilator support, brain ultrasonographic findings, major anomaly, and CP by thorough review of medical records. The major anomaly category included heart anomaly, central nervous system anomaly, chromosomal anomaly, specific syndromes, and complicated cleft palate. CP was identified based on the diagnosis by the pediatric or rehabilitation departments during the infant's follow-up period. We also searched for all cases of neonates born with normal cord pH (≥7.1) during the same study period who were diagnosed with CP during their follow-up period by screening the International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) infant codes (G80; CP) in our electronic medical system. Case confirmation was based on the clinical diagnosis made by pediatric or rehabilitation doctors as documented in the medical record. Statistical analysis was done using Mann-Whitney U tests, analysis of variance tests, Student's t-tests, and linear-by-linear association tests using SPSS software (version 25.0; IBM Corp., Armonk, NY, USA).

Table 1 shows a comparison of the clinical characteristics according to cord blood pH at birth. The low cord pH group was characterized by lower maternal height and weight and a higher rate of primiparity. The low cord pH group was associated with a higher grade of meconium staining and higher number of nuchal cord compared to the control group. In terms of the mode of delivery, the low cord pH group showed a significantly lower rate of elective caesarean section and higher rates of emergent caesarean section and vacuum-assisted delivery compared to the control group. Overall, rates of low Apgar scores at 1 minute (<4) and 5 minutes (<7) were 0.6% (n=79) and 0.4% (n=58), respectively. The low cord pH group was also associated with a higher rate of low Apgar scores at 1 minute and 5 minutes and NICU admission compared to the control group. Of note, the low cord pH group showed a lower mean neonatal weight (3.08 [1.68, 4.62] kg vs. 3.24 [1.51, 5.19] kg; P<0.001) and a higher rate of SGA compared to the control group (22.1% vs. 10.8%; P<0.001).

In our study population of term singleton pregnancies, the mean cord blood pH was 7.28±0.06, ranged 6.75–7.53 and the mean BE was −3.63±2.69. Rates of cord pH <7.2, <7.1, and <7.0 were 7.1% (n=1,011), 1.1% (n=163), and 0.3% (n=38), respectively. Next, we analyzed the mean value and distribution of cord blood pH according to the grade of meconium staining, number of nuchal cord, and mode of delivery (Table 2). There were significant differences in the mean cord pH based on the grade of meconium staining, number of nuchal cord, and mode of delivery (P<0.001 for all), but all mean values were within the normal range. Severe grades of meconium staining were associated with higher rates of cord pH <7.2, <7.1, and <7.0. The incidence of nuchal cord was also associated with a higher rate of cord pH <7.2 and <7.1 but not with cord pH <7.0. There were also significant differences in the rates of cord pH <7.2, <7.1, and <7.0 according to the mode of delivery.

Fig. 2 shows the clinical course information in cases with low cord blood pH. Among 163 cases, 49 babies (30.1%) were admitted to the NICU and 18 (11.0%) needed ventilator support. Ultrasonography of the brain was performed in 28.8% (n=47) of cases, and significant abnormal findings were observed in 27.7% (n=13) as follows: diffuse brain edema (n=7), grade 2 intraventricular hemorrhage (n=2), basal ganglia vasculopathy (n=2), Dandy–Walker complex (n=1), and myelination disorder (n=1).

Table 3 presents detailed clinical information on 3 cases of presumed CP with low cord pH. The baby in the first case was born at 39.0 weeks by prompt emergent caesarean section due to a non-reassuring FHR pattern detected at the time of admission. The neonatal weight was 3.234 kg and the meconium staining grade was 4+. The Apgar scores at 1 minute and 5 minutes were 0 and 4, respectively, but increased to 7 at 10 minutes. Cord blood pH was 7.021, and BE was −15.00 mmol/L. The baby was admitted into the NICU immediately and placed on a ventilator. An ultrasound examination of the brain on the first day of life revealed profound HIE findings, and the baby was diagnosed with moderate-severe cerebral coordination disturbance 1 month later. The baby was discharged 23 days after birth and, unfortunately, was found dead at home 6 months later. The second case was born by emergent caesarean section due to a non-reassuring FHR pattern developed during the second stage. The neonatal weight was 3.392 kg, and meconium staining grade was 1+. The Apgar scores at 1 minute and 5 minutes were 2 and 5, respectively. Cord blood pH was 6.887, and the BE was −18.30 mmol/L. The baby was admitted to the NICU and placed on a ventilator. Magnetic resonance imaging (MRI) of the brain checked on the 9th day showed a selective deep gray or white matter injury probably due to profound asphyxia. The electroencephalogram finding was also suggestive of diffuse cerebral dysfunction. The baby was lost to follow-up but was presumed to have CP. The third baby was born by spontaneous vaginal delivery without any remarkable event in the perinatal period. The baby had routine nursery care and was discharged uneventfully. However, after 4 months, the baby visited the pediatric clinic in our hospital due to difficulty with head tilting. Unfortunately, the baby suffered severe developmental delays and was finally diagnosed with ataxic CP at age 2.5 years, which was not HIE-related CP.

Lastly, we summarized the perinatal characteristics of the CP cases (n=18) with normal cord pH (≥7.1) at birth during the same study period (Table 4). Among 18 cases of CP, there were 4 cases with brain anomalies (hemimegalencephaly, severe band heterotopia, semilobar to lobar holoprosencephaly, and schizencephaly), 3 cases with specific syndromes (Schwartz-Jampel syndrome, Lennox-Gastaut syndrome, and Goldenhar syndrome), 3 cases of old brain infarction, and 1 case with a cardiac anomaly. Overall, SGA was complicated in 6 cases. Of note, the cord pH and BE were all within normal ranges and only 38.9% (n=7/18) of cases had NICU admission during the neonatal period.

Our data provided the following clinical information on a large hospital-based patient population over an 11-year study period. First, the mean cord blood pH was 7.28±0.06 among term singleton pregnancies. The rates of cord pH <7.2, <7.1, and <7.0 were 7.1%, 1.1%, and 0.3%, respectively. Secondly, low cord pH (<7.1) was associated with multiple antepartum factors including lower maternal height and weight, a higher rate of primiparity, lower neonatal body weight, and a higher rate of SGA. As for intrapartum factors, the grade of meconium staining, number of nuchal cord, and mode of delivery were associated with low cord pH, though the actual positive rate for low cord pH in the presence of these risk factors was quite low. Thirdly, among cases with low cord pH (<7.1), 30.1% were admitted into the NICU, 11.0% required ventilator support, and 28.8% warranted brain ultrasonography, with significant abnormal findings observed in 27.7% of cases. Lastly, the rate of presumable CP among cases with low cord pH was approximated at 1.8% (3/163), with 2 cases of spastic CP and 1 case of ataxic CP.

Although it is well documented that an umbilical artery pH <7.0 is considered one of the prerequisite criteria for a subsequent diagnosis of HIE [14], some studies on cord pH and FHR monitoring have chosen to use an umbilical artery pH of <7.1 as the critical acid-base level in an effort to identify the predictive risk factors at the stage before injury occurs [151617]. Similarly, as we aimed to investigate the neonatal outcomes in cases near the critical pH level, we defined “low” cord pH as <7.1 for research purposes. In our study, the mean cord blood pH and BE were 7.28±0.06 and −3.63±2.69 mmol/L, respectively, in term singleton pregnancies, and pathologic fetal acidemia with an umbilical artery pH <7.0 occurred in only 0.3% of cases. In our study population, the NICU admission rate was 6%. These results were quite similar to a previous study from Canada including more than 20,000 singleton infants born at term, in which the mean umbilical artery pH and BE were 7.24±0.07 and −5.6±3.0 mmol/L, respectively, and pathologic fetal acidemia (pH <7.0) occurred in 0.4% of cases with a NICU admission rate of 7.6% [18].

Multiple risk factors, both antepartum and intrapartum, have been reported to be associated with fetal acidemia. Antepartum risk factors include prior caesarean delivery, maternal age 35 years or older, prior neonatal death, single civil status, and short stature [19]. Intrapartum risk factors include the use of oxytocin, tachysystole, meperidine use, uterine rupture, cord complications, clinical indication of chorioamnionitis, meconium-stained amniotic fluid, breech presentation, operative delivery, and abnormal FHR during labor [10192021222324]. Our data also confirmed the antepartum factors of lower maternal height and weight, primiparity, lower neonatal birth weight, and SGA as risk factors for low cord pH.

Many studies have investigated the association between meconium staining and possible fetal acidemia, with conflicting results [2526]. One study reported a poor correlation between meconium-stained amniotic fluid and acute acidemia [27]. According to another study that included 323 term pregnancies with meconium-stained amniotic fluid, there was no correlation between umbilical cord acid-base measurements and fetal condition at birth [28]. In contrast, other studies have demonstrated that a severe grade of meconium staining is associated with cord blood pH [329]. Our study also confirmed that a high meconium staining grade was associated with lower umbilical cord pH. However, the actual frequency of low cord pH (<7.1) was low. In detail, the rate of low cord pH (<7.1) occurred in only 2.87% of cases of meconium staining grade +1 and 16.67% of cases of grade +4. Nuchal cord events increase with gestational age and are found in nearly 25% of deliveries at term [30]. While a nuchal cord is not associated with adverse outcomes in general, a multiple nuchal cord entanglement may be associated with a risk of meconium, abnormal FHR pattern, the need for operative vaginal delivery, mild umbilical artery acidemia at birth, and lower Apgar scores [3132]. Our data also identified an incidence of nuchal cord about 30% in term singleton pregnancies and showed that a multiple nuchal cord entanglement was associated with low cord pH, as expected. The mode of delivery is a well-known factor associated with low cord pH. Our data showed that the rate of low cord pH was higher in vacuum-assisted delivery than in any other mode of delivery, reflecting clinical situations requiring prompt delivery in cases of fetal jeopardy. However, the actual incidence of low cord pH (<7.1) was only 2.84%, while that of significant acidemia (<7.0) was exceedingly low (0.65%) in cases with vacuum-assisted delivery.

A population-based study indicated that the incidence of NE was 3.0 per 1,000 live births (95% confidence interval [CI], 2.7–3.3) and that of HIE was 1.5 per 1,000 live births (95% CI, 1.3–1.7) [33]. According to our study, the incidence of significant acidemia (<7.0) was 0.3% and that of umbilical cord pH <7.1 was 1.1%. Our observation that 1.8% of cases with low cord pH (<7.1) were finally presumed to have CP was similar to that of a recent observational cohort study including 51,519 term singleton neonates in which only 2.96% of neonates with cord pH <7.0 had adverse neurologic outcomes, including encephalopathy with seizures and/or death [34]. In their study, the authors concluded that most neonates with neurological morbidity have normal cord pH values, which is consistent with our data showing that most infants with neurodevelopmental disability during the same study period had normal umbilical cord pH levels. Collectively, these population-based studies suggest that variables other than acidemia are influencing the neurologic outcomes of infants.

Our study showed that an ultrasound of the brain was performed in only about one-third (28.8%) of cases with low cord pH, among which an abnormal finding was observed in 27.7%. In fact, diverse findings were identified, including diffuse brain edema, grade 2 intraventricular hemorrhage, basal ganglia vasculopathy, Dandy-Walker complex, and myelination disorder, which reflects the various underlying pathologies among cases of low cord pH. According to an MRI study of children with CP from a cohort of 334,339 infants at ≥36 weeks of gestation, it was demonstrated that only 5% had imaging abnormalities compatible with hypoxic-ischemic brain injury [35]. Other abnormalities included focal arterial infarction (22%), brain malformation (14%), periventricular white matter abnormalities (12%), generalized brain atrophy (7%), intracranial hemorrhage (5%), and delayed myelination (2%). Of note, a normal brain was seen in almost one-third of infants who underwent neuroimaging (31%).

The limitations of our study can be summarized as follows. First, considering the characteristics of a tertiary center where term neonates without NICU admission may not have long-term follow-up, we were unable to obtain complete information on the long-term outcomes of all neonates in this hospital population, especially those with cord pH >7.1, which may have resulted in an underestimation of the total prevalence of CP. Importantly, to overcome this limitation, we tried to search for CP cases among all the infants delivered in our hospital based on ICD-10-CM codes in our electronic medical record system, and we found 18 cases during the same study period that were diagnosed (or presumed) with CP later in life. The average follow-up period for 181 cases (163 cases with low cord pH and 18 cases of CP with normal cord pH) was 3.7 years. As a result, there were 21 CP cases, including 3 cases with low cord pH (2 cases of HIE-related and 1 case of ataxic CP), among 14,221 term singleton deliveries. Collectively, our data indicate that the total prevalence of CP in term pregnancies can be estimated at 1.48 per 1,000 live births, of which HIE-associated CP accounts for 9.5% (2 out of 21). This proportion seems to be similar to our previous study with a different study period (from October 1994 to December 2004) in which perinatal HIE events were evident in 20% of term CP cases [36]. Second, given the relatively higher proportion of high-risk pregnancies in our population, as reflected by the 6% NICU admission rate in term singleton pregnancies, generalizability to low risk population may be also limited. Thirdly, although vanishing twin was one of the criteria for exclusion from our study population, there is a chance that not all vanishing twin cases were ruled out in our population since clinical information in the first trimester may be not completely collected.

One of the strengths of our study is that we included a large number of consecutive deliveries in which the performance rate of cord ABGA was over 90%, thereby precluding intrinsic selection bias from case-control studies. In addition, we presented the objective association between umbilical cord pH and certain intrapartum factors including meconium staining, which is often overestimated in medicolegal disputes. Our data showing a very low incidence of HIE-associated CP even in neonates with low cord pH (1.23%, 2 out of 163) and a relatively high proportion of normal cord pH among total CP cases strongly support the expert opinion that if comprehensive etiologic evaluation is not possible, the term HIE should be replaced by NE, since neither hypoxia nor ischemia can be assumed to have been the unique causal mechanism [14].