Prediabetes in pregnancy – follow-up, treatment, and outcomes compared to overt pregestational diabetes

Abstract Objective There are limited data on follow-up, treatment, and maternal and fetal outcomes in women with prediabetes before or at the beginning of pregnancy. The aim of this study was to comprehensively characterize women with prediabetes compared to women with type 2 diabetes mellitus. Study design This was a retrospective cohort data from a single medical center treating women with pregestational prediabetes mellitus (PDM). Women were compared to pregestational overt type 2 diabetes mellitus (T2DM). Results Data were collected from 120 women in the PDM group and 86 women in the T2DM group. Baseline characteristics were comparable, albeit women in the PDM group arrived at medical attention significantly later, 55% after 15 weeks gestation. Women with PDM needed significantly less treatment to achieve glycemic control and glycated hemoglobin remained lower throughout pregnancy. Maternal and fetal outcomes were similar between groups, although significantly higher rates of macrosomia and neonatal jaundice were observed in the T2DM group. Conclusions The lack of clear guidelines causes a delay in the first prenatal visit of women with PDM. Comparable pregnancy outcomes may tip the balance toward acceptance of early treatment. Establishing clear guidelines will enable primary caregivers to refer prediabetic women sooner for lifestyle modifications and treatment if needed.


Introduction
Prediabetes poses an increased risk for the development of diabetes and cardiovascular disease throughout life [1]. A recent meta-analysis by Cai et al. found an increased risk of all-cause mortality and cardiovascular disease [2], stressing the importance of diabetes and prediabetes prevention. According to the American Diabetes Association (ADA) guidelines, diagnosis of prediabetes is based on fasting plasma glucose (FPG) between 100 and 125 mg/dL or glycated hemoglobin (HbA1c) between 5.7 and 6.4% or a twohour plasma glucose between 140 and 199 mg/dL after a 75-g oral glucose tolerance test (OGTT) [1].
If glucose values exceed these cutoffs, overt diabetes is diagnosed. Several prevention studies estimated the effectiveness of various interventions in delaying diabetes development [3][4][5][6].
Although the importance of prediabetes in the nonpregnant population is well established, the significance prior to, during, and after pregnancy on maternal and fetal outcomes is less clear. Currently, there is no uniform consensus on whether and when to intervene upon diagnosis. It is important to note that the term early gestational diabetes mellitus (GDM) encompasses women with prediabetes and relates to broader abnormal glucose values at the beginning of pregnancy.
The metabolic derangement in prediabetes can be regarded as an interim between GDM and overt diabetes. Data that compared normoglycemic women to GDM and overt diabetes found a significant increase in adverse pregnancy outcomes in the diabetic women [7,8]. However, most available data regarding prediabetes compared various outcomes to normoglycemic women or women with GDM, and not diabetic women and relied mostly on HbA1c [9][10][11].
Conflicting results emerge between studies comparing maternal and fetal outcomes in prediabetic women and women with normal glucose levels above and below certain HbA1c cutoffs. Studies that used an HbA1c cutoff of 5.7% did not demonstrate a difference in outcomes [9][10][11] while two other studies found adverse pregnancy outcomes [12,13] when a cutoff of 5.9% was employed. The authors demonstrated an increased risk of congenital anomalies, preeclampsia, shoulder dystocia, perinatal death [12], and macrosomia [13]. Recent ADA guidelines endorsed this cutoff [1]. In a more recent study by Mañ e et al. [14], an HbA1c of 5.8% was already associated with an increased risk of macrosomia, and first-trimester fasting plasma glucose (FPG) was not a better predictor for pregnancy complications.
Currently in Israel, women are screened in the first prenatal visit for FPG. In some cases, when a pregestational component is suspected, an early 100 g OGTT and/or HbA1c are performed. At our center, women diagnosed with prediabetes are referred to the Diabetes in Pregnancy Clinic, as are women with overt diabetes. They are followed and treated similarly by the same multidisciplinary team.
The aim of this study was to comprehensively characterize women with prediabetes and compare followup, treatment, and pregnancy outcomes to women with overt diabetes.

Study population
This retrospective cohort study included all women attending a multidisciplinary dedicated Diabetes in Pregnancy Clinic between 1 January 2014 and 31 December 2021 who were 18 years old and carried a singleton pregnancy. If a woman was pregnant more than once during the study period, only the first pregnancy was included in the analysis. We excluded women with GDM, preexisting type 1 diabetes and women that were lost to follow-up or delivered at another center. The multidisciplinary team includes an endocrinologist, maternal-fetal specialist, nurse, and dietician. Data were extracted from electronic medical records. The study protocol was approved by the local Institutional Review Board.
Women were divided to the prediabetes group (PDM) that included women with FPG of 100-125 mg/dL or HbA1c of 5.8-6.4% within three months prior to pregnancy to 15 weeks gestation without known preexisting diabetes and to preexisting type 2 diabetes mellitus (T2DM) group that included women with known T2DM or first diagnosed during pregnancy according to ADA guidelines.

Study procedures
All women with PDM and T2DM received lifestyle guidance with diet, physical activity, and weight gain recommendations. Blood glucose targets were 95 mg/dL for fasting and pre-prandial and 140 one hour post-prandial. If more than 30% of blood glucose measurements were above target during a one to two weeks period, drug therapy was initiated to achieve glucose targets of 90 mg/dL while fasting and 130 mg/dL one hour postprandial. First-line treatment initiated is insulin; however, if insulin was declined, metformin was offered. Metformin was continued if initiated before pregnancy or if deemed necessary to achieve glycemic control as add-on to insulin. Insulin pumps and continuous glucose monitors (CGMs) are reimbursed for women with T2DM but not for PDM; therefore, their use in PDM was very limited.
All women were screened after the first clinic visit for target organ damage including a fundoscopic exam, kidney function tests including microalbumin to creatinine ratio and neuropathy. Women were seen every one to four weeks as clinically indicated. Clinic personal could be contacted in between visits if necessary.
Data collected included demographics, anthropometric data, past medical history, laboratory results, treatment given, and maternal and fetal outcomes. Neonatal outcomes included gestational age at delivery, preterm birth before 37 weeks of gestation, birthweight, macrosomia (defined as birthweight 4 kg), large for gestational age (LGA) (defined as >90th percentile), small for gestational age (SGA) (defined as birthweight <10th percentile), neonatal hypoglycemia and hyperbilirubinemia (according to diagnosis in the medical chart), shoulder dystocia, respiratory distress syndrome, neonatal intensive care unit (NICU) admission within 24 h of birth, fetal abnormalities and intrauterine fetal death. Birthweight, macrosomia, LGA, and SGA were determined for the entire study cohort.
All decisions regarding the timing and mode of delivery were made by a maternal-fetal specialist according to local protocol.
All women with PDM received a recommendation to complete blood tests for either a 75-g OGTT or FPG and HbA1c three months following delivery, and results were recorded. Follow-up continued until six months post-partum. If a woman performed more than one test during that time, the worst was recorded.

Statistical analysis
Categorical and nominal variables were reported by prevalence and percentage, and continuous variables were reported as mean and standard deviation (SD) or as median and range. Continuous variables were tested for normality by the Shapiro-Wilk test. Normally distributed variables were compared between groups by Student's t-test and abnormally distributed variables by the Mann-Whitney U-test.
Comparisons of proportions were analyzed by Pearson's Chi-square (v 2 ) test or Fisher's exact test. Analyses were performed with SPSS27 (SPSS Inc., Chicago, IL). A p value of <.05 was considered statistically significant.

Study population characteristics
During the study period, 120 women with PDM and 86 women with T2DM met the inclusion criteria. Among the PDM group, 57 women (48%) were diagnosed based on impaired FPG alone, five (4%) by HbA1c 5.8% alone, and 58 women (48%) had both.
Patient characteristics are presented in Table 1. Groups were comparable for age, body mass index (BMI), obesity rates, smoking at the beginning of pregnancy, and obstetric history. About a fifth of the women in each group were nulliparous. Women with T2DM were significantly more likely to report a firstdegree family history of diabetes and to be diagnosed with polycystic ovary syndrome (PCOS). Women with T2DM presented to the clinic significantly earlier during gestation, with 70% presenting before 15 weeks gestation and only 6% after 24 weeks, while only 46% of the women with PDM presented before 15 weeks gestation and 30% after 24 weeks gestation (p<.001).

Target organ damage
A minority of women had microvascular complications in the two groups (Table 1). One woman with T2DM had non-proliferative diabetic retinopathy and a second had hypertensive retinopathy, both remained stable throughout pregnancy. Significantly more women with T2DM had nephropathy at first visit. None had diabetic neuropathy. One woman in the PDM group had a history of mild stroke related to known hypercoagulability.
Glycemic control during follow-up (Supplementary Table 1) Women with PDM visited the clinic significantly less than those with T2DM (6 ± 3 vs. 8 ± 4 visits per pregnancy, p<.001). Per definition, glucose values and HbA1c were significantly lower in pre-pregnancy and in the first trimester in women with PDM compared to T2DM. HbA1c remained significantly lower at the second and third trimesters in the PDM group although glycemic control was good in both groups. Table 1) A minority of women were controlled with diet only in the PDM and T2DM groups (8.3% and 3.5%, respectively, p¼.245). The rest were treated with insulin, metformin, or both. Women with T2DM received twice as much metformin to achieve glycemic control compared to PDM (56% vs. 23.3%, p¼.001). Women with T2DM were significantly more likely to receive multiple daily injections (MDIs) (24% vs. 10%, p¼.01), and to use insulin pumps and CGMs. Thus, overall treatment burden was higher in women with T2DM compared to PDM. In the PDM group, treatment needs were stratified according to diagnosis with impaired FPG alone, impaired HbA1c alone, or both. The latter group received metformin trifold (36% vs. 11%, p¼.003) compared to impaired FPG or HbA1c alone. No other differences were noted between the groups regarding treatment (data not presented).

Maternal and fetal outcomes
Maternal and fetal outcomes are listed in Table 2. No differences were noted in preeclampsia rates, mean birthweight, LGA, SGA, gestational age at delivery, preterm births and primary or secondary cesarean delivery rates. There was a significantly higher rate of macrosomia in the T2DM group and more newborns that were diagnosed with neonatal jaundice. The rate of admissions to the NICU was comparable between groups.

Post-partum follow-up
Of the 120 women with PDM, 101 (84%) were tested for glycemic status until six months post-partum. Of those, 27 (26.7%) done a 75 g OGTT, 96 (95%) FPG and 89 (88.1%) had an HbA1c result. In 23 women glucose levels returned to normal, 60 (59.4%) remained in the prediabetic range, and 18 (17.8%) were compatible with overt diabetes. Women with both impaired FPG and abnormal HbA1c did not differ from those with abnormal FPG or HbA1c alone.

Discussion
In the current study, we comprehensively characterized pregnancy outcomes of women with PDM compared to T2DM. Most baseline characteristics were comparable between the two groups, including age, obesity rates, gestational weight gain, and obstetric status. Women with T2DM had higher rates of positive family history for diabetes, a diagnosis of PCOS and diabetic nephropathy. Importantly, although prediabetes was already known or diagnosed in the first trimester, the first clinic visit occurred significantly later in the PDM group and 55% had their first visit after 15 weeks of gestation. Lack of awareness and consensus regarding the need for prompt treatment may explain this observation. Notably, only a minority of women in both groups achieved glycemic control with diet only, and the vast majority needed medical treatment. However, although initial BMI and gestational weight gain were similar in both groups, women with PDM needed significantly less treatment to achieve glycemic control, and HbA1c remained lower throughout pregnancy. The higher treatment burden in T2DM women is expected due to a more significant initial hyperglycemia and insulin resistance. Rowan et al. [15] demonstrated that over 90% of women with an HbA1c 5.9% needed pharmacotherapy either with metformin or insulin in addition to lifestyle adjustments. In our cohort, a similar percentage needed pharmacotherapy, although our cutoff was 5.8%.
Although early treatment is not uniformly recommended in women with PDM, one of the largest studies to date by Sweeting et al. suggests that it should be strongly considered [16]. In this study, WHO/IADPSG cutoffs were used for early GDM diagnosis. The authors stratified 4873 women across the entire spectrum of hyperglycemia. Sixty-five women had preexisting T2DM, 68 had GDM diagnosed before 12 weeks' gestation, 1247 had GDM diagnosed between 12 and 23 weeks' gestation, collectively called early GDM and 3493 had GDM diagnosed after 24 weeks' gestation. Women were treated regardless of the timing of diagnosis as soon as it was clinically indicated. They concluded that despite intensive treatment regimens, the results demonstrate a continuum of risk for adverse maternal outcomes according to the type and timing of diabetes diagnosis.
Surprisingly, in our cohort, although treatment burden was lower in the PDM group and HbA1c was also lower throughout pregnancy, most maternal and fetal outcomes were comparable between groups. Significant differences were noted for macrosomia and neonatal jaundice in the T2DM group. Though all women received treatment according to best-known practice, results suggest that there is still a need to improve outcomes. Advocacies for delaying treatment include the lack of definite efficacy, risk of overtreatment leading to increased risk of SGA, cost, emotional burden, and an increased risk for interventions. In our cohort, SGA and cesarean delivery rates were comparable between groups.
Several studies demonstrate a high rate of adverse events in women with prediabetes. In the prospective cohort by Hughes et al. [12], women were screened for HbA1c before 20 weeks' gestation. Out of 16,122 women, 200 were found with HbA1c between 5.9 and 6.4%. When analyzing pregnancy outcomes, the authors excluded women treated for GDM to avoid potential confounding factors. They compared untreated women with screening HbA1c 5.9% and <5.9% and found that the former had poorer pregnancy outcomes. Two recent studies also demonstrate an increased risk of LGA in prediabetic women. In a prospective study by Quansah et al. [17], 76 women with PDM as diagnosed by ADA prediabetes criteria before 20 weeks' gestation were compared to 1185 women with GDM. They had similar maternal and fetal outcomes except for higher rates of LGA (25% vs. 15%, p¼.02). Jayasinghe et al. [18] prospectively examined 2709 women until 13 weeks' gestation for early hyperglycemia using both WHO/IADPSG criteria and ADA criteria. Overall, prediabetes was diagnosed in 68 women, and 21 were diagnosed with overt diabetes, and the relative risk for LGA was 1.45 and 2.3, respectively.
Two randomized controlled trials attempted to solve the dilemma of whether to treat early or wait. The first by Osmundson et al. [19] examined 83 women enrolled before 14 weeks' gestation with HbA1c between 5.7 and 6.4%. Usual care was compared to early treatment. The primary endpoint was a 75-g OGTT at 26-28 weeks. Secondary outcomes included cesarean delivery, birthweight, weight gain, and HbA1c change. Although treated women had significantly lower HbA1c over time, this did not translate to a significant difference in any of the secondary parameters. It was acknowledged that a difference may exist, but the study was underpowered to detect it.
The second study by Hughes et al. [20] examined 49 women enrolled before 14 weeks' gestation with HbA1c between 5.9 and 6.4%. Women were randomized to early intervention vs. standard of care. However, recruitment rates were lower than expected, and non-adherence to the allocated treatment protocol was significant (42% in the early intervention group and 30% in controls). Overall, numbers were too small to reach significance. Taken together, most data support early treatment in PDM; however, large randomized clinical trials are needed to validate this.
All women in the PDM group were asked to complete glycemic evaluation after birth. Of the 84% that complied, less than a quarter reverted to normal glucose post-delivery. The rest remained prediabetic or progressed to overt diabetes, highlighting the need for close follow-up. Most available studies did not continue follow-up post-partum. In the study by Quansah et al. [17], the post-partum prevalence of metabolic syndrome, prediabetes, and diabetes were 1.8, 3.1, and 7.4 times higher than GDM.
The strengths of our study include a well-characterized cohort of women followed for the entire pregnancy by the same multidisciplinary team with elaborated data available before, throughout, and after pregnancy. We captured both FPG and HbA1c data when available. Limitations include a relatively small sample size that was underpowered to detect differences in complications and clinical outcomes. Furthermore, we do not have long-term follow-up data on the children born to these mothers.
To conclude, prediabetic women are a high-risk group during and after pregnancy. Indeed, these women are at increased risk for diabetes and need careful follow-up post-partum. Although the need for early treatment in pregnancy is still under debate, as well as the cutoff for treatment, we believe correct labeling, recognition, and referral are indicated as soon as possible. The similar rate of complications compared to T2DM and lack of a harmful effect in our study, along with the accumulating evidence in the literature, may tip the balance toward acceptance of early treatment. Establishing clear guidelines will enable primary caregivers to refer prediabetic women sooner for lifestyle modifications and treatment if needed.