Well-managed gestational diabetes mellitus is no longer a risk factor for neonatal RDS — a case-control study

BACKGROUND Diabetes in pregnancy used to be considered associated with a higher risk of respiratory distress syndrome(RDS) in neonates. However, whether well-managed gestational diabetes remains an independent risk factor is unclear. This study was to determine the associations of them. METHOD This was a case-control study conducted at the Obstetrics & Gynecology Hospital of Fudan University, Shanghai, China. We collected data from 101 RDS infants and 101 RDS infants from among 1749 infants, through a standardized protocol with predened inclusion and exclusion criteria. Two investigators enrolled in recording same datas twice to minimize bias. Assessment of diabetes management was based on HbA1c and random blood glucose measurements. Univariable and multivariable logistic regression was performed to calculate the odds ratio (OR). An intergroup analysis was conducted between RDS infants and those without RDS, and a subgroup analysis was conducted between RDS neonates born to women with gestational diabetes and those born to women without gestational diabetes.


Background
Respiratory distress syndrome (RDS) is one of the most common and severe respiratory diseases in neonates. Although RDS most occurs in early preterm babies (GA lower than 34 weeks), RDS still was founded in a signi cant number of late preterm infants (GA between 34 weeks and 36 weeks and 6 days) and term infants [1].Due to the increased complications and higher mortality that they experience, these patients are noteworthy as well. Gestational diabetes mellitus (GDM) was considered associated with an increased risk of neonatal complications, including the higher incidence of RDS [2,3].Some previous studies have demonstrated that RDS occurs more frequently in infants whose mothers have GDM [4,5] due to hyperinsulinemia induced by hyperglycemia. Hyperinsulinemia appears to delay surfactant synthesis by interfering with lung maturation [6][7][8]. However, a recent secondary analysis study based on the results of a randomized controlled trial suggested that GDM was not a risk factor for RDS [9].
Additionally, another study focusing on preterm infants (22-33 weeks gestation) reached a similar conclusion [10]. These contradictions didn't reveal whether a well managed GDM is still increase the risk of neonatal RDS,and whether nowdays therapy for GDM is suitable. Hence, we conducted a case-control study containing 101 RDS infants and 101 non-RDS infants to test the hypothesis that well-managed gestational diabetes does not increase and exacerbate RDS in late preterm or term infants.

Study subjects
This is a unmatched case-control study which comprised 202 neonates with a GA was higher than 34 weeks who were admitted to the neonatal intensive care unit (NICU) at the Obstetrics & Gynecology Hospital of Fudan University, Shanghai, China, from 1st June 2017 to 1st June 2019. All datas were collected anonymously medical record system of Obstetrics and Gynecology Hospital of Fudan University and collection was authorized. We excluded infants with severe cardiac malformation and incomplete antenatal record data, as well as infants who died during hospitalization, from the analysis. The case group comprised 101 infants with radiographically con rmed RDS (further criteria below). The control group consisted of 101 non-RDS infants who were admitted with a diagnosis of transient tachypnea, asphyxia, congenital pneumonia, hyperbilirubinemia, or low birth weight. We randomly selected controls from the neonates admitted to the same ward over the same time period as with the case group; identical inclusion and exclusion criteria were applied. Randomization was based on a random number table linked to the patient identi cation number of each infant.

Diagnostic criteria for RDS, GDM, and PGDM
The oral glucose tolerance test (OGTT) was used to diagnose GDM and PGDM at the Obstetrics & Gynecology Hospital of Fudan University. Positive results for fasting, one-hour postprandial, and two-hour postprandial blood glucose concentrations were levels higher than 5.1, 10.0, and 8.5 mmol/L, respectively [11]. The diagnostic criteria for RDS included (1) a clinical picture of infants with the onset of progressive respiratory failure shortly after birth with respiratory failure that manifested by an increase in the work of breathing and an increase in the oxygen requirement; (2) a characteristic chest radiograph featuring low volume, the classic diffuse ground-glass appearance and air bronchograms. De nition of well-managed GDM[12-14]: Glycated hemoglobin (HbA1c) was lower than 6.5%, or the concentrations of fasting, one-hour postprandial, and two-hour postprandial blood glucose were all lower than 5.3 mmol/L, 7.8 mmol/L, and 6.7 mmol/L, respectively. Additionally, random blood glucose (RBG) < 11.1 mmol/L was considered a reference standard for well-managed diabetes.
De nition of confounders: Maternal: (1)pregnancy induced hypertension(PIH) was de ned as systolic blood pressure ≥ 140 mmHg and/or diastolic blood pressure ≥ 90 mmHg) at ≥ 20 weeks of gestation in the absence of proteinuria or new signs of end-organ dysfunction [15]. (2)Premature rupture of fetal membranes(PROM) means fetal membranes break 8 hours or earlier before delivery. (3)Maternal prenatal fever was de ned as body temperature is above 38.0℃ whenever before, during or after delivery. (4)Infection and potential infection:including septicemia, intrapartum fever of mother (T > 38.0℃), congenital pneumonia. (5)Placental abnormality includes placental abruption, velamentous insertion, battledore placenta, placental implantation, placenta previa, and placental adhesion. (6)Umbilical cord abnormality refers to prolapse, presentation, tieing of umbilical cord and cord around the neck; (7)The criteria of asphyxia, which is our routine suggested by Newborn Professional Committee of Chinese Medical Association, is (a)Apgar score is less than 7 at 1 min or 5 min after birth and (b)umbilical artery serum pH is blow 7.15 [19]. (8)Anemia was de ned as hemoglobin of umbilical cord blood is below 17.0 g/dL for term infants and 16.0 g/dL for preterm infants.

Statistical analysis
We summed the data as frequencies or percentages for categorical variables and as the means and standard deviations or medians and interquartile ranges for continuous variables, according to the distribution. Differences between the groups were compared by the chi-square or Fisher's exact test for categorical variables, Student's t-test, or Mann-Whitney U test for continuous variables, depending on the distribution. Binary logistic regression was used to assess associations of GDM or PGDM with RDS; the model was rst unadjusted and then adjusted for GA, asphyxia, and other potential risk factors. A nominal 2-sided probability value < 0.05 was considered to indicate statistical signi cance. All of the calculations were performed using SPSS 23.0 (SPSS Inc. Chicago, IL).

Results
We reviewed the medical records of patients admitted to the neonatal intensive care unit (NICU) at the Obstetrics & Gynecology Hospital of Fudan University. After excluding 33 neonates according to the exclusion criteria, 1446 late preterm and term infants were eligible for analysis. All 101 infants who developed RDS were allocated to the case group. For the control group, 101 neonates were selected randomly from the 1345 other neonates with non-RDS diseases. An intergroup comparison and a subgroup comparison were performed (Fig. 1).
Clinical data: The clinical characteristics of the case group and control group are presented in Table 1. The preterm birth rate and low birth weight rate were both signi cantly different between the two groups (P < 0.05) after adjusting for maternal age, the occurrence of elderly pregnancy, PIH, cesarean rate, sex, gestational age, and birth weight. SGA, AGA, LGA was categorised individually according to the birth weight below the 10th centile,between 10th and 90th centile, and above 90th . † Elderly pregnancy:maternal age is grater than 35 years. ‡ Infection and potential infection:including septicemia, intrapartum fever of mother (T > 38.0℃), congenital pneumonia.
* there is a signi cant difference (P < 0.05, in chi-square test) We analyzed the HbA1c and RBG levels of mothers with diabetes in the two groups to determine whether their GDM was well-managed. Table 2 and Fig. 1 illustrate the diabetes features of those with GDM and PGDM in the two groups diabetes features, including morbidity, treatment, rate of tests (HbA1c, RBG), and mean gestational age before diagnosis, were similar for both groups ( Table 2). Levels of HbA1c at diabetes was diagnosed, HbA1c right beofre delivery and RBG before delivery had no signi cant difference, and all of them were below a well-controlled level. These results indicated that both groups of diabetic mothers were well managed (Fig. 2). OGTT: Oral glucose tolerance test; RBG: random blood glucose; GA: gestational age * there is a signi cant difference (P < 0.05, in t-student test).
Univariable and multivariable logistic regression was performed to con rm whether well-controlled diabetes (GDM and PGDM) was still a risk factor for RDS in late preterm and term infants (Table 3). After adjusting for confounders, the incidence of diabetes showed no signi cant differences between the cases and the controls. This means that well-controlled diabetes was no longer a risk factor for these babies. To investigate whether babies born to mothers with gestational diabetes had poorer outcomes than those born to mothers without gestational diabetes, we grouped RDS neonates into two subgroups: the nondiabetic group and the diabetic group, and a subgroup analysis was conducted between these two groups. Table 4 displays the RDS characteristics of these neonates. The transfer rate of the diabetic group was signi cantly lower than that in the non-diabetic group, indicating that the number of severe cases in the diabetic group was less than that in the non-diabetic group. And for mild and moderate patients, who didn't need transfer for further and more advanced treatment, the total dose of PS applications, incidence of complications, and respiratory support rates (total and separate) ( Table 4) as well as different respiratory support duration (Fig. 3) showed no signi cant differences between the two groups. All of these results showed that situation of RDS neotates did not deteriorate because of GDM and PGDM. Table 5 shows that there was a higher usage rate of hood oxygen, while the usage rates of other respiratory supports were similar. As hood oxygen was often utilized before RDS was diagnosed, gestational diabetes might have led to a more prolonged latent period of RDS.

Discussion
This study has three clinically relevant ndings: (1) The similarity in morbidity, severity, and outcomes between infants born to mothers with GDM and infants born to mothers without GDM revealed that wellmanaged gestational diabetes did not increase the incidence of RDS in late preterm or term infants. (2) For RDS infants, GDM did not exacerbate their situation including complications, need for pulmonary surfactant(frequency and dose), and need for respiratory support(frequency and duration There are potential reasons for this observation. First, diabetes of the mother may cause fetal hyperglycemia and later hyperinsulinemia, which works as an inhibitor of the key enzyme of lipid synthesis [6,26] and production of the vital protein SP-B [27]. Second, preterm births were signi cantly more common in women with diabetes, which directly led to immature fetal lungs and a lack of pulmonary surfactant [28,29]. This preterm risk seems to be higher for pregestational diabetes type I than for type II [30]. In addition, macrosomia, which is more likely to result in cesarean section (CS), was more prevalent in babies with diabetic mothers [29,31], and CS is related to an increased risk of RDS [32]. A study based on a pregnancy cohort with 110879 women con rmed this point. It was found that GDM was associated with a 1.63 RR of RDS, and a 1.40 RR of neonate hypoglycemia [33], which means that infants were in uenced by hyperinsulinemia before delivery. There are a small number of papers that studied the effect of different types and severity of gestational diabetes on RDS. Francisco et al. reported that women with diet-controlled GDM and normoglycemic women had infants with appropriate lung maturation, while the insulin-treated group had more immature lungs [34]. This implied that worse islet cell function of the mother drove more noticeable adverse effects on the fetal lung. However, none of these studies have detailed glycemic and HbA1c levels during pregnancy to illustrate whether the con icting results may be due to poor diabetes management.
Strengths and Limitations of our study: To our knowledge, this is the rst study concerned with the details and validity of diabetes control during pregnancy. To shed light on this problem, we focused on diabetes indicators at different times: HbA1c and OGTT at diagnosis and HbA1c and random blood glucose right before delivery. These results not only con rmed that diabetes was well-managed in our study but also indicated that an HbA1c of 6.5% is a safe threshold for gestational diabetes to avoid additional RDS risk (Fig. 2). As far as the reliability of the data was concerned, although not every pregnant woman had their diabetes monitoring recorded due to limitations of the retrospective study, most of them (13/17 to 15/16) ( Table 2) provide credible results. With a standardized protocol to collect all of the data, information bias was minimized.
Another advantage is that we elaborated on the effect of diabetes on RDS infant outcomes in more detail for several different aspects, including counts and doses of PS usage, complications, and the duration of respiratory support. The relationship between different types respiratory support and gestational diabetes con rmed that well-managed diabetes did not aggravate RDS, even if the duration was somewhat discrete. This point was also supported indirectly by comparisons of other complications, such as PPHN, pneumonia, or pneumothorax. However, what hampered us in obtaining a rmer conclusion was the low morbidity of nonpreterm RDS, resulting in the small number of neonates in the case group. This also means that additional research, which includes a larger sample size and premature infants, is needed to illustrate whether GDM has a different effect on various gestation ages. There are some minor limitations in our study. For example, limited by time and labor cost, it was di cult to collect information on all 1345 contemporary infants without RDS for the control group. We used a simple randomization to select and

Conclusion
Well-managed gestational diabetes did not increase the incidence of RDS in late preterm or term infants.
And for those infants with RDS, GDM or PGDM of their mothers did not exacerbate their complications, need for pulmonary surfactant, or frequency and duration of respiratory support use. These results indicated that current GDM management targets including (1)FBG below 11.1 mmol/L, (2)HbA1c was lower than 6.5%, and(3)OGTT results below 5.3 mmol/L, 7.8 mmol/L, and 6.7 mmol/L are enough for reducing risk of neonatal RDS. Besides, this study also implied that management methods of diet and exercise are su cient to reach this GDM management target. Additionally, our research showed some intra-uterine or extra-uterine hypoxic-ischemic factors, such as asphyxia, placental abnormality, and fetal distress, are likely to relate to neonatal RDS.This needs some more basic researches to illustrate the mechanism.
review approximately 1/10 of the records, which may have resulted in a sampling error. In addition, as an obstetrics and gynecology hospital, our NICU lacks advanced equipment and surgical collaborators to rescue critical patients; we must transfer these patients to specialized pediatric hospitals. The absence of their detailed information forced us to conclude that they had a worse prognosis, which may have introduced bias in the comparison of complications. However, this bias does not change the conclusion that diabetes did not increases the risk of RDS.
On account of the detailed examination in this study, there is an interesting nding in our work: RDS resulted in an increased usage rate of hood oxygen, which is the primary respiratory support used in TTN or for the initial stage of RDS in our routine procedures. This means that GDM can prolong the latent period of RDS, or some RDS was possibly derive from TTN. It should be noted that the most appropriate comparison was between infants born to mothers with wellmanaged GDM and infants born to mothers with poorly managed GDM. Nevertheless, there were very few pregnant women with poorly managed GDM in Shanghai, China, due to improved antenatal examinations. Therefore, our study was a case-control study designed to compare RDS infants and non-RDS infants and examine the well-managed GDM incidence of their mother to determine the odds ratio (OR). Availability of data and materials The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.All datas were collected from database of medical record system of Obstetrics and Gynecology Hospital of Fudan University and collection was authorized.
Competing interests There is no con ict of interest associated with this manuscript.  Flow diagram of participants.

Figure 2
The comparison of laboratory test results for diabetes indicators between neonates with and without RDS. A. OGTT results for diabetes diagnosis, B. HbA1C results at diabetes diagnosis, C. HbA1C right before delivery, D. RBG right before delivery. NS: there is no signi cant difference; P>0.05, which was calculated with Student's t-test for variables with a normal distribution.

Figure 3
Comparison of different durations of respiratory support between babies born to mothers with gestational diabetes and babies born to mothers without gestational diabetes. A, B, C, D present separate comparisons for treatment durations of hood oxygen, CPAP, MV, and total respiratory support, respectively, between infants born to mothers with gestational diabetes and infants born to mothers without gestational diabetes. NS: there was no signi cant difference; P>0.05, which was calculated with the Mann-Whitney U test for a nonnormal distribution.