What Is the Cost of Reducing Adverse Pregnancy Outcomes in Patients With GDM: A Retrospective Cohort Study

Luiza Oleszczuk-Modzelewska 2nd Department of Obstetrics and Gynecology, Medical University of Warsaw, 00-315 Warsaw, 2 Karowa St Aneta Malinowska-Polubiec (  anetapolubiec@interia.eu ) 2nd Department of Obstetrics and Gynecology, Medical University of Warsaw, 00-315 Warsaw, 2 Karowa St Ewa Romejko-Wolniewicz 2nd Department of Obstetrics and Gynecology, Medical University of Warsaw, 00-315 Warsaw, 2 Karowa St Agnieszka Zawiejska Department of Medical Simulation, Chair of Medical Education, Poznan University of Medical Sciences, 60-512 Poznan, 41 Jackowskiego St Krzysztof Czajkowski 2nd Department of Obstetrics and Gynecology, Medical University of Warsaw, 00-315 Warsaw, 2 Karowa St

performed before the 24th or after the 28th week of pregnancy, or with incomplete results of the threepoint OGTT.
In Poland, the 2011 criteria for diagnosing diabetes considered an abnormal result to be at least one of the following glucose values in the 75-g OGTT: fasting ≥ 100 mg/dL, 1 hr ≥ 180 mg/dL, or 2 hr ≥ 140 mg/dL. Meanwhile, the criteria adopted in 2014, according to IADPSG and WHO, considered at least one of the following results of the 75-g OGTT to be abnormal: fasting ≥ 92 mg/dL, 1 hr ≥ 180 mg/dL, or 2 hr ≥ 153 mg/dL.
This study compared patients with gestational diabetes to a group of women in whom GDM was excluded. The risk factors for obstetric complications were the age of the pregnant woman, the value of the maternal body mass index (BMI) before pregnancy, gestational weight gain (GWG), the circumference of the mother's abdomen measured at the level of the navel before delivery, women who had had at least one child in the past (multiparity), birth body weights of older children over 4000 g, fasting blood glucose, blood glucose in the second hour of the 75-g OGTT, diagnosis of diabetes in the current pregnancy, the need to implement insulin therapy to maintain normoglycaemia and female sex of the foetus.
The impact was analysed of the aforementioned factors on the following maternal and paediatric obstetric outcomes: weight gain of pregnant women with respect to Institute of Medicine (IOM) recommendations, incidence of pregnancy-induced hypertension or pre-eclampsia, incidence of composite adverse perinatal outcomes, including haemorrhage or perinatal trauma, the necessity to terminate pregnancy by caesarean section due to foetal indications, the birth weight of new-borns, including the percentage of children with macrosomia, the incidence of hypoglycaemia or hyperbilirubinemia (treated with phototherapy) in the rst three days of the new-borns' life, and the percentage of children with congenital anomalies or rare neonatal complications.
Descriptive results are expressed as the mean ± standard deviation or median (interquartile range). Categorical variables are expressed as percentages.
Multivariate logistic regression (forward method) was used to identify predictors for dichotomous adverse maternal and foetal outcomes in the pooled analysis of the entire group (patients with GDM and normoglycaemic controls). Variables that correlated with the outcomes with a p < 0.1 in the bivariate analysis were included in the models. The results are presented as adjusted odds ratios (aORs) and 95% con dence intervals (95% CIs). Predictors for continuous variables were identi ed using multiple linear regression models (forward method), with gestational weight gain or birth weight as dependent variables.
All variables that showed a bivariate correlation with a p < 0.1 with these outcomes were entered in the models as independent variables. A p < 0.05 was considered statistically signi cant in the multivariate analysis.
A ROC analysis was used to calculate the diagnostic power of maternal abdominal circumference measured before delivery as a predictor of adverse neonatal outcomes. Two-sided p < 0.05 was considered statistically signi cant.
Ethics approval for this study was obtained from the Warsaw Medical University institutional review board (AKBE13/15).

Results
In the analysed material, based on the OGTT, in the group of 285 (58.5%) pregnant women, at least one of the glycaemic values met the criteria for a diagnosis of diabetes, either according to the 2011 or 2014 criteria, while the other group of 202 (41.5%) women did not meet any of the criteria for diagnosing diabetes.
Women with at least one abnormal glycaemic result, regardless of the adopted criteria for the diagnosis of diabetes, were compared to the group in which the OGTT result was normal, regardless of the criteria and had the following characteristics: signi cantly older mean age (32.4 ± 4.7 vs 31.1 ± 4.1, p < 0.001) with signi cantly higher BMI before pregnancy (24.6 ± 4.6 vs 22.7 ± 3.6, p < 0.01), the vast majority of patients were obese (13.5% vs 3%, p < 0.001), less likely to give birth for the rst time (38.2 vs 50.2%, p < 0.03), more often had relatives with diabetes (48.1% vs 23.8%, p < 0.01) and had a history of GDMcomplicated pregnancy (18.8% vs 0%, p < 0.001). These results are shown in Table 1. If we compared patients with GDM to the normoglycaemic control group, no statistically signi cant differences were observed with respect to the incidence of maternal complications, the condition of newborns or in the incidence of neonatal complications between the groups. There was only a trend for borderline statistical signi cance of more frequent urgent caesarean section due to foetal indications in the group with gestational diabetes (26.4% vs 16.2%, p = 0.06). These results are shown in Table 2.   Table 4 shows data on maternal abdominal circumference measured before delivery as a predictor of obstetric complications in the entire cohort (GDM patients and the control group). The size of the maternal abdominal circumference before delivery was a strong factor correlating with the occurrence of perinatal complications in both the mother and the foetus in the entire examined cohort. A circumference over 100 cm increased the risk of at least one pregnancy complication (increased blood loss, soft tissue injuries, pre-eclampsia) by almost 40% (OR 1.38, p < 0.001). A circumference over 98 cm increased the risk of foetal macrosomia by 20% (OR 1.24, p < 0.005), and a circumference over 104 cm increased the risk of one of the complications during the neonatal period by 50% (OR 1.54, p < 0.005). Furthermore, in the group with gestational diabetes, a circumference over 103 cm doubled the risk of neonatal hypoglycaemia during the rst days after delivery (OR 2.01, p < 0.0001).

Discussion
The number of patients with gestational diabetes in the world is continuously increasing. According to various estimates, over the last twenty years, the percentage of women with GDM has increased by 10-100%, especially in highly developed countries, and in 2019, hyperglycaemia was diagnosed in approximately 16% of pregnancies worldwide, of which GDM accounted for 84% of all cases [16][17][18][19][20][21]. This fact allows us to predict a signi cant increase in the number of obstetric complications and forces researchers to identify factors that may affect their development.
In our study, we demonstrated that the diagnosis of diabetes in pregnancy increases the risk of having a child with macrosomia by 10-fold (OR 10.4, p < 0.005) and 13-fold in multiparous women (OR 13.9, p < 0.005). These results are identical to other available publications [22][23][24][25]. To date, the individual in uence of blood glucose values at individual measurement points in the 75-g OGTT on obstetric complications is not fully understood. The HAPO study, which was the basis for changing the existing criteria for the diagnosis of GDM, demonstrated a linear relationship between maternal glucose levels and the child's birth weight [10]. Zhu et al. and Zawiejska et al. found that macrosomia was diagnosed signi cantly more often in children of patients with fasting hyperglycaemia [19,26]. On the other hand, Kerenyi et al. [25] found that the curve illustrating the relationship between fasting glucose measured during the 75-g OGTT and the birth weight of the foetus and the risk of LGA was U-shaped (p = 0.004), indicating that both in patients with low and high fasting blood glucose levels, the risk of foetal hypertrophy was increased. In a publication by Black et al. [27], attention was also drawn to the signi cant in uence of hyperglycaemia in the 2nd hour of the 75-g OGTT on the increased risk of pregnancy induced hypertension (PIH), preterm labour and hyperbilirubinaemia in new-borns.
In our study, we did not identify any correlation between the glycaemic status of patients from particular groups and the percentage of maternal (here: pre-eclampsia) or foetal complications ( . Similar relationships were demonstrated in the work of Bodnar et al. [35]. In our study, we found that higher pre-pregnancy BMI values correlated with a higher risk of pre-eclampsia during pregnancy (OR 1.15, p < 0.001). On the other hand, the increased abdominal circumference measured in patients before delivery had a signi cant impact on increasing risk of perinatal complications in women (increased blood loss, injury of soft tissues of the birth canal) [OR 1.08, p < 001], caesarean section due to urgent indications connected to the risk to the foetus (8% increase) and high birth weight of new-borns (12-fold increase in the risk; OR 12.1, p < 0.001). Moreover, among other complications during the early neonatal period, we observed in uence of high maternal abdominal circumference on increasing risk of hypoglycaemia in the rst days of life (by 12%; OR 1.12, p < 0.001), the need for phototherapy due to hyperbilirubinaemia (increase by 7%) and the risk of at least one complication during the neonatal period, i.e., hyperbilirubinaemia, hypoglycaemia or respiratory disorders (by 8%; OR 1.08, p < 0.001). Additionally, Gao et al. showed that both overweight or obesity before pregnancy and increased abdominal circumference in a patient signi cantly increased the risk of developing GDM, caesarean section delivery and macrosomia [36]. It is worth emphasizing that in the population of non-pregnant women, waist circumference is considered an indicator of insulin resistance, and its increased value has been included in the criteria for diagnosing metabolic syndrome [37][38][39][40][41].
Some sources report waist-to-hip ratio (WHR) to be superior to BMI for predicting the risk of developing type 2 diabetes, hypertension and cardiovascular disease in adults [42][43][44][45][46]. Obviously, the measurement of abdominal circumference in pregnancy is a speci c measurement that is technically di cult and related to a non-standard population, but the relationships observed in our study between neonatal complications and increased abdominal circumference in term pregnancy con rm that this parameter also informs the "metabolic condition" of the mother and should be taken into account in the context of expected perinatal complications. Our observation of the in uence of maternal parameters related to insulin resistance and an excessive percentage of adipose tissue on the risk of obstetric complications may also explain the persistence of a high percentage of obstetric complications in the population of pregnant women with hyperglycaemia in pregnancy, despite optimization of metabolic control. This was also con rmed by data from our multivariate regression models, which indicate that weight gain in pregnancy or the circumference of the pregnant woman's abdomen measured before delivery, and not the severity of carbohydrate tolerance disorders, remain risk factors for signi cant obstetric complications.
This means that the parameters describing the "maternal metabolic status" remain a signi cant risk factor for adverse maternal-foetal outcomes when effective treatment eliminates the risk associated with hyperglycaemia in pregnancy.  [49]. Similar results were found in the work of Papazian et al. [50]. In our study, we did not identify any signi cant relationships between weight gain in pregnant women and the risk of obstetric complications. We did notice, however, that patients with GDM, compared to healthy pregnant women, had a signi cantly higher pre-pregnancy BMI (24.6 ± 4.6 vs 22.7 ± 3.6, p < 0.01) and were mostly obese (13.5% vs 3%, p < 0.001). It is also worth emphasizing that the diagnosis of diabetes in pregnancy was associated with a twice lower risk of excessive weight gain in pregnancy (OR 2.43, p = 0.001), although the increased risk of excessive weight gain in pregnancy was signi cantly associated with a higher BMI before pregnancy (OR 1.12, p < 0.001), higher fasting glucose in the OGTT (1.03, p < 0.05) and lower blood glucose in the 2nd hour of OGTT (OR 0.99, p < 0.05). Fasting hyperglycaemia is a marker of hepatic insulin resistance and one of the components of metabolic syndrome [51]. Therefore, in the context of our research, the positive relationship between excessive weight gain in pregnancy and fasting hyperglycaemia should be interpreted as a clinical manifestation of the relationship between pregnancy weight gain and insulin resistance.
As we calculated, the diagnosis of diabetes during pregnancy reduced the risk of excessive weight gain in pregnancy by 30% (OR 0.31, p < 0.001) and increased the risk of weight gain under 12 kg (OR 4.44, p < 0.01) by four-fold. These results also suggest that treatment of hyperglycaemia with well-controlled diabetes in pregnancies that still experience complications result from non-glycaemic risk factors, including components of metabolic syndrome. One of the effects of multidisciplinary care for pregnant women with GDM may be low weight gain in pregnancy, which is di cult to interpret unequivocally due to the lack of guidelines for the group of patients with gestational diabetes. For others, it is possible to slow down intrauterine growth, which protects against macrosomia and intrauterine death of the foetus.
In light of the data available to us, it seems that the price of these bene ts may be an increased risk of accelerated weight gain in infants and obesity in early school age children [52,53]. However, one should also take into account the latest data presented by the LifeCycle Project consortium, which showed, in a population of approximately 200,000 pregnant women, that in women with a BMI > 30, the optimal weight gain for reducing obstetric complications is lower (0-6 kg) than that recommended by the IOM for pregnant women with a similar BMI (5-9 kg) [34]. Additionally, authors of a retrospective observational study of 2,842 women with GDM published in 2020 con rmed the dominant pattern of weight gain in pregnancy below the level recommended by IOM in this population (50.3% of the examined patients) [54].
The results of our study indicate that the glycaemic status of patients may be a predictor of certain maternal complications, including abnormal weight gain in pregnancy, but neither gestational diabetes nor blood glucose levels at individual 75-g OGTT measurement points were predictors of neonatal complications in the study cohort. On the other hand, it is noteworthy that the anthropometric conditions of pregnant women and gestational weight gain, which are indicators of the "metabolic status" of women, may signi cantly correlate with the occurrence of obstetric complications, although their longterm effects on the mother and child require further study.

Conclusions
The results of our research indicate another area of medical intervention in pregnancy complicated by high metabolic risk. Apart from normalization of glycaemia, stabilization of the percentage of adipose tissue and non-glycaemic obstetric risk factors may also be necessary to obtain further improvement in obstetric outcomes in this pregnant population.

Declarations
Ethics approval and consent to participate