Introduction

Type 1 diabetes in the mother is associated with high risk of adverse neonatal outcomes. The most common complications are congenital malformations, premature delivery, fetal macrosomia, and birth injuries.1,2 In a study by Mimouni et al3, over 25% of infants delivered to diabetic women had perinatal asphyxia. Nowadays, due to significant improvements in diabetic and obstetric care, these numbers seem to have been reduced.4 Attempts should be made to prevent neonatal acidosis, because it is associated with a higher risk of postnatal adverse outcomes including neurological injuries, multiple organ dysfunctions, and increased long-term morbidity and mortality rates.5,6

The analysis of umbilical artery pH values has been widely adopted in clinical practice. It is especially recommended for high-risk populations such as pregnant women with diabetes. Some factors in the general population, such as prior cesarean delivery, maternal age of 35 years or older, previous neonatal death, general anesthesia, a thick meconium, uterine rupture, and abnormal fetal heart rate, have been identified as risk factors for neonatal acidosis.7 In women with type 1 diabetes, antepartum hyperglycemia correlates with umbilical erythropoietin levels at birth, which suggests that maternal hyperglycemia induces fetal hyperinsulinemia, hypoxemia, and, as a consequence, neonatal acidosis. However, there are limited data on the determinants of neonatal acidosis in pregnancy complicated by type 1 diabetes. We hypothesized that insufficient control of maternal diabetes might be a dominant factor predisposing to neonatal hypoxemia. Therefore, the aim of this study was to identify maternal factors associated with low cord arterial pH values in newborns of mothers with type 1 diabetes.

Patients and methods

Clinical and laboratory data on 1069 women with type 1 diabetes and their infants for a period between 1993 and 2015 were extracted from the clinical registry of a tertiary care center of the Poznan University of Medical Sciences (Poznań, Poland). The hospital, including its outpatient clinic, is the biggest perinatal center for pregnant women with diabetes in Poland. It provides care for patients from the Wielkopolska Province (population of approximately 3.4 million people). A total of 280 women with diabetes were excluded from the analysis due to miscarriage (pregnancy loss before 22 weeks’ gestation), major birth defects, multifetal pregnancy, diabetes other than type 1 or unspecified types, or incomplete information. Finally, data from 789 women with type 1 diabetes were included in the study. All women received intensive insulin therapy during pregnancy.

Registry of pregnant women with diabetes

Clinical and laboratory data of women with type 1 diabetes have been collected and stored in our center continuously since 1993. According to the Polish Diabetes Association recommendations and our internal standards, every woman with preexisting type 1 diabetes from our region is immediately referred to our department once pregnancy is confirmed. Women with diabetes without complications have at least 3 planned, short-stay hospital admissions during pregnancy: in the first trimester, at midpregnancy (20–24 weeks’ gestation), and near delivery (34–39 weeks’ gestation). Patients who require more vigilant surveillance are admitted more frequently. Between hospital admissions, patients are referred for regular check-ups every 2 weeks at a hospital-based outpatient clinic.8,9 We recommend that all of our pregnant patients with type 1 diabetes deliver in our center. In this way, we have direct access to their pregnancy and peripartum data, including the umbilical cord blood gas analysis, which is routinely performed in our center.

Umbilical cord blood gas analysis

For the purpose of this study, we analyzed pH values in the umbilical artery. Umbilical cord blood was collected separately from the artery and the vein, soon after birth, from a doubly clamped segment of the cord, and then immediately transported to the laboratory for analysis. During the sampling procedure, it is sometimes difficult to differentiate the artery from the vein, so we assumed that samples with lower pH came from the artery. Based on the pH values of the umbilical arteries of their infants, women were divided into 2 groups: normal pH, defined as pH of 7.1 or higher, and low pH, defined as pH lower than 7.1.

Statistical analyses

Statistical analyses were performed using the MedCalc software for Windows, version 12.1.3.0 (MedCalc Software, Mariakerke, Belgium). Testing for normality of data distribution was performed using the D’Agostino–Pearson test. The t test was used to measure the significance of the difference between 2 continuous variables when data fitted a normal distribution, with results presented as mean (SD). In the case of nonnormally distributed data, comparisons were made using the Mann–Whitney test, with results presented as the median and interquartile range. The χ2 test was used for the comparison of categorical variables. A logistic regression analysis was used to identify the determinants of low pH in the umbilical artery, with data presented as odds ratios (ORs) and 95% CIs. Both the univariate and multivariate models were built to search for possible associations between variables. A 2-tailed P value of less than 0.05 was considered significant. The presence of vasculopathy was defined as being diagnosed with at least 1 of the following: retinopathy, nephropathy, or ischemic heart disease. Small for gestational age (SGA) was defined as a birth weight lower than the 10th percentile, and large for gestational age (LGA) was defined as a birth weight greater than the 90th percentile, using age- and sex-specific regional growth charts.

The degree of prematurity was described based on gestational age, and was divided into subgroups as follows: extremely preterm, less than 28 weeks’ gestation; very preterm, 32 to 28 weeks’ gestation; and moderate to late preterm, 32 to 37 weeks’ gestation.

Ethical approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee as well as with the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards.

This study is a retrospective analysis of routinely collected data and therefore is exempted from the formal review by the Poznan University of Medical Sciences Ethics Committee.

Results

The characteristics of the study subgroups are shown in Table 1.

Table 1. Characteristics of women with type 1 diabetes divided into groups based on pH values in the umbilical arteries of their infants
ParameterNormal pH (≥7.10) n = 717Low pH (<7.10) n = 72P value

Maternal age, y

28 (5)

28 (5)

0.7

Maternal age at diagnosis of type 1 diabetes, y

15 (8)

15 (9)

0.8

Diabetes duration, y

11 (7)

11 (8)

0.8

BMI, kg/m2

23.2 (7.2)

23.9 (5.7)

0.2

HbA1c, %

I trimester

7.3 (1.7)

7.7 (1.6)

0.08

II trimester

6.1 (1.1)

6.4 (1.2)

0.02

III trimester

6.3 (0.9)

6.6 (1.1)

<0.01

Pregnancy planning, n (%)

226 (31.5)

22 (30.5)

0.9

Vasculopathy, n (%)

189 (26.3)

22 (30.5)

0.5

Gestational age at delivery, wk

37 (2)

37 (3)

0.46

Birth weight, g

3298 (777)

3231 (853)

0.49

Data are presented as mean (SD) unless otherwise indicated.

Abbreviations: BMI, body mass index; HbA1c, glycated hemoglobin A1c

Among the study subgroups, there were no differences in maternal age, age at onset of diabetes, diabetes duration, the proportion of women receiving preconception care, the proportion of women with vasculopathy, gestational age at delivery, and body mass index (BMI).

A total of 72 infants (9.1%) had low pH in the umbilical artery. Low pH values were associated with a decreased Apgar score at 1 minute (OR, 0.76; 95% CI, 0.70–0.82; P <0.001) and 5 minutes (OR, 0.77; 95% CI, 0.69–0.85; P <0.001) after birth. Maternal age (OR, 0.99; 95% CI, 0.94–1.04; P = 0.72), age at diagnosis of diabetes (OR, 1.01; 95% CI, 0.97–1.04; P = 0.75), and diabetes duration (OR, 0.99; 95% CI, 0.95–1.04; P = 0.78) were not associated with low umbilical artery pH. Maternal pre-pregnancy BMI (OR, 0.98; 95% CI, 0.94–1.02; P = 0.4), the presence of diabetic vascular complications (OR, 1.23; 95% CI, 0.73–2.01; P = 0.44), chronic hypertension (OR, 1.32; 95% CI, 0.65–2.70; P = 0.44), and gestational hypertension or preeclampsia (OR, 0.73; 95% CI, 0.32–1.65; P = 0.44) had no impact on low umbilical artery pH.

Low pH values were not associated with gestational age at delivery (OR, 0.95; 95% CI, 0.86–1.05; P = 0.34) or the degrees of prematurity, namely, late to moderate preterm (OR, 1.32; 95% CI, 0.76–2.31; P = 0.33), very preterm (OR, 1.55; 95% CI, 0.45–5.36; P = 0.49), and extremely preterm (there was only 1 extremely preterm infant with normal pH in the umbilical artery of 7.29). There was an association between maternal glycated hemoglobin (HbA1c) levels measured before delivery and low pH in the umbilical artery (OR, 1.40; 95% CI, 1.11–1.78; P <0.01). A similar association was found for HbA1c levels measured between 20 and 24 weeks’ gestation (OR, 1.29; 95% CI, 1.03–1.63; P = 0.03). There was no association between the first-trimester HbA1c level as well as lack of preconception care and low pH in the umbilical artery. Predictors of low pH in the umbilical artery based on logistic regression models are shown in Table 2.

Table 2. Predictors of low pH in the umbilical artery based on logistic regression models
Predictors of low pHOdds ratio (95% CI)P value

HbA1c II trimestera, %

1.29 (1.03–1.63)

0.02

HbA1c III trimestera, %

1.40 (1.11–1.78)

0.03

Urgent cesarean sectionb, yes/no

1.64 (1.11–2.44)

0.01

a Univariate logistic regression

b Model adjusted for HbA1c levels in the third trimester

Abbreviations: see Table 1

Of the 107 newborns delivered via urgent cesarean section, 14 of them (13.1%) had low pH in the umbilical artery. There were 27 vacuum-assisted deliveries, and 3 of those newborns (11.1%) had low pH in the umbilical artery. There were 4 forceps deliveries, with low umbilical artery pH found in 2 newborns (50%). A logistic regression revealed an association between the urgent cesarean section and low pH in the umbilical artery (OR, 1.64; 95% CI, 1.11–2.44; P = 0.01), and this association was independent of HbA1c levels measured before delivery.

No association between either neonatal LGA or SGA and low pH in the umbilical artery was observed.

To address the possible influence of typical obstetric factors on umbilical artery pH values, we performed a separate analysis excluding urgent cesarean and operative vaginal deliveries. The association between HbA1c levels measured between 20 to 24 weeks’ gestation and before delivery remained significant (OR, 1.36; 95% CI, 1.04–1.77; P = 0.02 and OR, 1.41; 95% CI, 1.07–1.87; P = 0.01; respectively). Similarly, no association was found between the HbA1c level in the first trimester and umbilical artery pH (OR, 1.16; 95% CI, 0.99–1.36; P = 0.07). There was still no association between low umbilical artery pH and maternal age, age at diagnosis of diabetes, diabetes duration, maternal prepregnancy BMI, the presence of diabetic vascular complications, chronic hypertension, gestational hypertension or preeclampsia, and neonatal LGA or SGA.

Discussion

This study is among the first to evaluate determinants of neonatal acidosis in women with type 1 diabetes in a cohort of this size.

In all women, we assessed multiple factors that might have potentially influenced neonatal acidosis. However, we demonstrated that only maternal levels of HbA1c in the second and third trimesters of pregnancy and urgent cesarean sections were associated with low pH in the umbilical artery. In a population-based study of pregnant women with type 1 diabetes, Klemetti et al10 demonstrated an association between HbA1c values measured closest to delivery and several adverse neonatal outcomes, including asphyxia, which is in line with our results. However, the authors did not find a relationship between HbA1c levels measured in the first and second trimesters and neonatal asphyxia. In our study, there was also no association between the levels of HbA1c in the first trimester and low pH in the umbilical artery. This might be explained by the fact that in general, patients’ metabolic control was improved during pregnancy because of the intensification of treatment after admission to our center. Similarly, Striakov et al11 revealed that in patients with preexisting diabetes mellitus, increased levels of HbA1c during early pregnancy are not useful in predicting most adverse outcomes. Cahill et al12 investigated the impact of glycemic control on neonatal morbidity in pregnancies complicated by gestational and pregestational diabetes and showed that the level of HbA1c in late third trimester is the most useful predictor for neonatal risk in women with diabetes, which is in line with our findings. Cahill et al12 focused mainly on Apgar scores of less than 7 at 5 minutes, but Locatelli et al5 demonstrated that, as expected, Apgar scores correlate with neonatal acidemia.

Elevated BMI predisposes women to numerous complications during pregnancy and labor. Cnattingius et al13 showed that a BMI value greater than 25 kg/m2 in pregnant patients with diabetes is associated with an increased risk of low Apgar scores and severe asphyxia. In our study, we did not observe an association between maternal BMI and low pH value in the umbilical artery, but we focused only on pregnant women with type 1 diabetes, who are rarely obese or overweight.

Diabetic vasculopathy is a well-known risk factor for unfavorable pregnancy outcomes. Type 1 diabetic women with vasculopathy are at higher risk of preeclampsia and pathological fetal growth, as shown by Howarth et al.14 However, the authors did not investigate pH values in the umbilical artery. In our study, maternal vasculopathy was not associated with low pH in the umbilical artery. It may be explained by the fact that the majority of women with vasculopathy had elective cesarean delivery due to vasculopathy or other maternal indications.

It is essential that women with diabetes plan their pregnancies and achieve adequate metabolic control from the very beginning of pregnancy. Previous studies showed that patients with pregestational diabetes who plan their pregnancies achieve better obstetric outcomes.15,16 In our cohort, one-third of women attended preconception counseling, and these proportions were similar in groups with normal and low pH. This suggests that even in women with poorly controlled diabetes, intensification of treatment and normalization of glycemic control may have beneficial effects on the neonatal condition at birth.

Vascular complications of diabetes, such as nephropathy and retinopathy, are associated with longer duration of diabetes and with lack of adequate metabolic disease control. In 1949, White17 classified diabetes in pregnancy according to such factors as the age at onset of the disease, diabetes duration, and the presence of diabetic complications. Based on this classification, Klemetti et al18 investigated the pregnancy outcomes in women with type 1 diabetes. The authors found that the levels of HbA1c in the first trimester, preeclampsia rates, cesarean delivery rates, deliveries before 37 weeks’ gestation, and neonatal intensive care unit admissions increased from class B to F. It seems that the duration of diabetes can influence these outcomes. In our study, the duration of diabetes or the presence of maternal vasculopathy were not associated with an increased risk of low pH in the umbilical artery.

Miailhe et al19 demonstrated that the HbA1c level of 6.4% or higher at delivery was associated with urgent prelabor cesarean deliveries due to abnormal nonstress test results. Although the false-positive rate for the abnormal nonstress test is relatively high and can reach 55%, in numerous cases it can be a manifestation of varying degrees of fetal acidosis.20 In our study, we used a direct method of evaluation of neonatal condition at birth, but our results seem to confirm the findings of Miailhe et al.19 Nonetheless, numerous urgent cesarean deliveries in our cohort could not be explained by maternal HbA1c levels and were associated with typical obstetric indications.

Study limitations

The study population included consecutive female patients with type 1 diabetes. These women were managed according to the same protocol in a single obstetric center for women with diabetes. However, because we collected data over a long period, we were not able to access all data concerning the delivery, such as the total duration and decision-to-delivery time in the cases of urgent cesarean and operative vaginal deliveries. Because all of these factors may influence umbilical artery pH, we conducted a subanalysis excluding women who delivered via urgent cesarean and operative vaginal procedures (vacuum, forceps). This left us with a subgroup of women who delivered vaginally or abdominally, with no signs of fetal distress. In this subgroup, the determinants of low umbilical artery pH remained unchanged. This suggests that the majority of women who delivered newborns with low umbilical artery pH did not present with any signs of fetal distress. This is an important clinical finding because fetal metabolic alterations, including acidosis, as an effect of maternal hyperglycemia may precede intrauterine fetal death.21-23

Another limitation of the study is that our records included only early neonatal data, and we were not able to analyze long-term development of the neonates. Umbilical artery pH is an easily obtainable and useful parameter expressing neonatal asphyxia. However, it may not always correlate with the long-term neonatal development.

Conclusions

Among modifiable risk factors, the lack of sufficient glycemic control in pregnancy is the strongest predictor of neonatal acidosis in women with type 1 diabetes. Therefore, improved glycemic control throughout pregnancy may reduce the risk of neonatal distress postpartum. However, a significant number of cesarean deliveries due to acute signs of fetal distress, resulting in low umbilical artery pH values in the infant, cannot be explained by maternal hyperglycemia, and these are probably related to classic obstetric factors.