Treatment of Gestational Diabetes Mellitus and Offspring Early Childhood Growth

Abstract

Background

Gestational diabetes mellitus (GDM) is associated with fetal overgrowth, and certain treatments are associated with an increased risk of macrosomia. However, there are limited data about the long-term effect of GDM treatment on childhood growth.

Methods

Cohort study of 816 women with GDM and their offspring delivered between 2009 and 2012. Childhood height and weight through age 3 were collected from the medical record and z-scores and body mass index (BMI) were calculated. We assessed the association between GDM treatment and childhood growth using linear mixed modeling.

Results

Treatment was divided into medical nutritional therapy (MNT) (n = 293), glyburide (n = 421), and insulin (n = 102). At delivery, birthweight, z-score, and BMI were higher in the offspring of women treated with either glyburide or insulin compared to MNT. However, weight, z-score, and BMI were similar among all offspring at 6 months and 1, 2, and 3 years of age. After controlling for covariates, there were differences in the weight z-score (P = 0.01) over the 3-year period by treatment group, but no differences in weight (P = 0.06) or change in BMI (P = 0.28). Pairwise comparisons indicated that insulin was associated with more weight gain compared with MNT (0.69 kg; 95% CI, 0.10-1.28; P = 0.02) and glyburide was associated with a trend toward lower weight z-score compared with MNT (−0.24; 95% CI, −0.47 to 0.003; P = 0.05).

Conclusion

Despite growth differences detected at birth, we observed no meaningful differences in childhood growth from 6 months to 3 years among treatment groups, including in the offspring of women with GDM treated with glyburide.

Gestational diabetes mellitus (GDM) affects approximately 7% of pregnancies in the United States, (1) and it is associated with increased risk for adverse pregnancy outcomes, including fetal overgrowth, neonatal morbidity, cesarean delivery, and hypertensive disorders of pregnancy (2, 3). GDM treatment decreases the risk for adverse pregnancy outcomes including macrosomia (4, 5). Professional organizations recommend a trial of medical nutritional therapy (MNT) followed by the addition of pharmacologic therapy as needed to achieve glucose control throughout gestation (6-8).

Past guidelines suggested an equal role for insulin and oral hypoglycemic agents (glyburide and metformin) for the treatment of GDM (9). However, recent updates have highlighted certain concerns about glyburide use (8). An early randomized clinical trial comparing glyburide to insulin for GDM treatment found that glycemic control and perinatal outcomes were similar, and that glyburide did not cross the placenta (10). As a result of this trial, glyburide therapy for GDM increased from 7.4% to 64.5% between 2000 and 2011 (11). However, more recent studies demonstrated that glyburide crossed the placenta, with cord blood levels that are approximately 70% of maternal serum levels and it may be associated with an increased risk for macrosomia and neonatal hypoglycemia (12-16). Importantly, these reports lacked data on maternal glycemic control and pre-pregnancy obesity, which are also independently linked to fetal overgrowth.

Furthermore, there are limited data on the long-term effects of different maternal GDM treatments on offspring growth. Offspring of women with GDM have increased risk for obesity, metabolic syndrome, diabetes, and adverse neurodevelopmental conditions that may be related to the intrauterine environment (17-20). Therefore, we sought to compare childhood growth through 3 years of age in a cohort of the offspring of women with GDM treated with either MNT, glyburide, or insulin.

Materials and Methods

We conducted a retrospective study of women with singleton gestations and GDM who were delivered at Magee-Womens Hospital (University of Pittsburgh, Pittsburgh, PA) from January 2009 to October 2012. Women were identified using the ICD-9 codes 648.01 (diabetes-delivered) and 648.81 (abnormal glucose tolerance-delivered). Derivation of the study cohort is shown in Fig. 1. Out of a total of 38 222 deliveries, we identified 2289 women with an ICD-9 code consistent with the diagnosis of diabetes. We excluded any second pregnancy that occurred within the study time frame, and we also excluded women who reported a diagnosis of pre-gestational diabetes at their first prenatal visit or if they had a first-trimester HbA1c value > 6.5%. Women with GDM were included only if their records were available for review and if they had either a 50-gram, 1-hour glucose challenge test (GCT) that exceeded 200 mg/dL as per institutional policy, or if they had 2 or more abnormal values on a 3-hour, 100-gram oral glucose tolerance test (OGTT) as defined by the Carpenter-Coustan criteria (21). We excluded 147 women because we did not have records of their GDM testing, 139 women who did not meet diagnostic criteria for GDM based on their testing results, and 35 women who were diagnosed with GDM by alternate methods, including either a 75-gram, 2-hour OGTT, or capillary blood glucose monitoring. Of the 1374 women who met the inclusion criteria, we identified 816 offspring whose pediatric records were available for review because they received ongoing care through the same health care system. Regulatory approval was obtained from the University of Pittsburgh Institutional Review Board, and informed consent was not required given the retrospective nature of the study.

Figure 1.

Derivation of the study cohort.

Women included in this study received prenatal care in the Maternal Fetal Medicine and Obstetrics clinics at our health system. Patients received nutritional counseling and education about recommended weight gain based on their pre-pregnancy BMI at a centralized office. All women were treated following an institutional guideline for the clinical management of GDM. Women underwent individualized MNT counseling by certified diabetes educators in either a group (approximately 75%) or individualized (approximately 25%) setting. Women were advised to follow a carbohydrate-controlled diet with approximately 40% to 50% of energy to be obtained from complex carbohydrates, 20% to 30% from protein, and 20% to 30% from fat. Self-monitoring of blood glucose 4 times daily was encouraged, and glucose targets included a fasting value less than 95 mg/dL, and 1-hour postprandial values less than 140 mg/dL (8, 22).

In order to assess maternal glycemic control, 7 days of consecutive self-monitored blood glucose were extracted from the medical record at 4-week intervals. Mean fasting and postprandial blood glucose values were calculated across gestation, with blood glucose data available for 690/816 (84.6%) women. Approximately 7 days after their diabetes education session, women were scheduled for follow-up with a Maternal Fetal Medicine (MFM) physician to discuss their GDM control and make recommendations regarding therapy. The choice of therapy (glyburide versus insulin) and decision regarding when to initiate or change therapy was determined by the MFM physician. Our primary exposure was GDM treatment, consisting of either MNT, glyburide, or insulin therapy at delivery.

Maternal pre-pregnancy body mass index (BMI) was calculated from the pre-pregnancy weight reported in the medical record, and the reported pre-pregnancy weight had a strong correlation with the measured weight at the first prenatal visit (r = 0.98, P < 0.001). Maternal pre-pregnancy overweight/obesity was defined using the WHO guidelines for classification of BMI. Gestational weight gain was defined as insufficient, sufficient, or excessive for each pre-pregnancy BMI category as defined in the Institute of Medicine 2009 guidelines (23).

Perinatal outcomes were abstracted from the maternal medical record and compared among the various treatment groups. Hypertensive disorders of pregnancy were considered together as a single outcome, and we included the following diagnoses: gestational hypertension, pre-eclampsia without severe features, pre-eclampsia with severe features, and superimposed pre-eclampsia. Other outcomes considered included cesarean delivery and preterm birth <37 weeks’ gestation.

Macrosomia was defined as birth weight >4000 grams, and large for gestational age (>90th percentile for gestational age) or small for gestational age (<10th percentile for gestational age) status based on US national birth weight data (24). Neonatal outcomes included neonatal intensive care unit (NICU) admission, hypoglycemia (defined as a glucose value less than 35 mg/dL within the first 24 hours of life), hyperbilirubinemia requiring phototherapy, respiratory distress syndrome, congenital anomalies, and neonatal death. We also defined a composite neonatal morbidity consisting of hypoglycemia, hyperbilirubinemia, or respiratory distress syndrome. Finally, we compared rates of congenital anomalies among groups.

Approximately 60% of the offspring from women who deliver at Magee-Womens Hospital continue to receive care through pediatric clinics affiliated with the University of Pittsburgh Medical Center. Women with and without offspring childhood follow-up data were similar with regards to maternal age, race, rates of private insurance, pre-pregnancy BMI, weight gain in pregnancy, weeks at GDM diagnosis, and type of diabetes therapy at delivery. Women with offspring childhood data available were less likely to smoke (8.1% vs 12.6%), and they had slightly lower mean postprandial blood glucose (123.1 ± 14.0 vs 124.9 ± 14.2 mg/dL). Of the offspring with follow-up data available, 171/816 (21.0%) had data at 1 time point, 150/816 (18.6%) had data at 2 time points, 192/816 (23.5%) had data at 3 time points, and 303/816 (37.1%) had data at 4 time points. Records were linked with delivery information when available, using the infant medical record number. We collected infant and childhood data at time points of 6 months and 1, 2, and 3 years, including measured length/height and weight. The closest measurements were included ± 4 weeks for the 6- and 12-month visits and ± 8 weeks for the 24- and 36-month visits. Weight was compared at each time point among groups, and we used the 2000 Centers for Disease Control and Prevention (CDC) Growth charts to calculate each child’s weight z-score, standardized to the reference population for the child’s age and sex (25). BMI was also calculated, and we categorized the BMI as either normal weight or overweight/obese using the CDC’s standard threshold of the 85th percentile for childhood overweight (26). The medical records were also reviewed to assess breastfeeding. For this analysis, breastfeeding was defined as any breastfeeding reported at an outpatient pediatric visit.

Statistical analyses were conducted using SAS Version 9.4 (SAS Institute Inc, Cary, NC). Distributions of variables were tested for normality using visual inspection of histograms and the Shapiro-Wilk W-test. Baseline demographics, delivery outcomes and infant/childhood growth were compared by GDM treatment at delivery using chi-squared statistics, Fisher exact test, or ANOVA as appropriate. The primary analysis utilized linear mixed modeling for offspring weight (unstandardized and standardized), weight z-score, or BMI as a function of time, treatment group, the interaction of time and treatment group, and a random subject effect. Additionally, we included as covariates baseline pregnancy characteristics that differed significantly between treatment groups, including maternal age, pre-pregnancy BMI, first-trimester systolic blood pressure, 50-gram GCT, number of weeks at time of GDM diagnosis, and nulliparity. Of primary interest was whether the change in outcome over 3 years was significantly different between treatment types, which was assessed using linear contrasts. If significant, pairwise differences were assessed between treatment types. As an exploratory measure, and due to the pattern of weight and BMI over time, we fit piecewise linear mixed models that allowed for separate slopes during the first 6 months and the subsequent 30 months.

Results

Infant and childhood follow-up information was available for 816/1374 (59.4%) of women with GDM. Among these 816 women, 293 (35.9%) were treated with MNT, 421 (51.6%) were treated with glyburide, and 102 (12.5%) were treated with insulin at delivery. The need for medical therapy was lower among women who had a fasting glucose value on their OGTT <95 mg/dL when compared with those who had a fasting glucose value ≥95 mg/dL (54% vs 80%; P < 0.001) Women in both the glyburide and insulin groups had a similar time from GDM diagnosis to medication initiation (3.3 ± 2.3 vs 2.8 ± 2.5 weeks; P = 0.06). Among the 102 women who were on insulin at delivery, 27 (26.5%) received glyburide prior to transition to insulin. Women who initially received glyburide and then transitioned to insulin had a shorter duration of glyburide exposure than women who were receiving glyburide at delivery (4.9 [interquartile range (IQR) 3.6-8.0] vs 7.1 [IQR 5.1-9.1]) weeks. The women who were transitioned from glyburide to insulin received insulin for a median of 5.4 (IQR 3.5-7.6) weeks. As shown in Table 1, women who were on glyburide or insulin at delivery were slightly older than women treated only with MNT. Women who required medical therapy for their GDM also had higher BMIs and both higher results on their glucose testing (GCT, OGTT) and higher mean fasting and postprandial glucose values (Table 1).

Table 1.

Demographic and Clinical Characteristics, Overall and by Treatment Group

	N if < 816	Total	MNT	Glyburide	Insulin	P value
		(N = 816)	(n = 293)	(n = 421)	(n = 102)
Maternal age (years)		31.6 ± 5.5	30.9 ± 5.5	31.9 ± 5.4	32.0 ± 5.4	0.04
Race						0.64
White		617 (75.6%)	220 (75.1%)	319 (75.8%)	78 (76.5%)
Black		118 (14.5%)	39 (13.3%)	62 (14.7%)	17 (16.7%)
Other		81 (9.9%)	34 (11.6%)	40 (9.5%)	7 (6.9%)
≥ Some college education		586 (71.8%)	215 (73.4%)	299 (71.0%)	72 (70.6%)	0.76
Private insurance		600 (73.5%)	214 (73.0%)	309 (73.4%)	77 (75.5%)	0.89
Nulliparous		414 (50.7%)	160 (54.6%)	219 (52.0%)	35 (34.3%)	0.002
Pre-pregnancy BMI (kg/m2)	806	30.0 ± 7.9	27.4 ± 7.0	31.1 ± 8.0	32.6 ± 8.4	< 0.001
BMI category	798					< 0.001
Normal		239 (29.9%)	115 (41.1%)	108 (25.9%)	16 (15.8%)
Overweight		190 (23.8%)	75 (26.8%)	91 (21.8%)	24 (23.8%)
Obese		369 (46.2%)	90 (32.1%)	218 (52.3%)	61 (60.4%)
Weight gain category	800					0.08
Under		185 (23.1%)	79 (27.5%)	87 (21.0%)	19 (19.4%)
At		273 (34.1%)	102 (35.5%)	141 (34.0%)	30 (30.6%)
Over		342 (42.8%)	106 (36.9%)	187 (45.1%)	49 (50.0%)
Tobacco use		66 (8.1%)	23 (7.8%)	36 (8.6%)	7 (6.9%)	0.84
CHTN		51 (6.3%)	14 (4.8%)	27 (6.4%)	10 (9.8%)	0.20
First-trimester systolic blood pressure	783	115.9 ± 11.3	114.3 ± 10.8	116.5 ± 11.5	118.0 ± 11.7	0.006
First-trimester diastolic blood pressure	783	71.7 ± 8.2	71.0 ± 7.9	71.9 ± 8.5	72.7 ± 7.7	0.19
50 g GCT	749	170.2 ± 29.0	163.0 ± 22.8	172.3 ± 30.1	182.7 ± 34.9	<0.001
100 g OGTT
Fasting	727	92.1 ± 16.3	85.1 ± 11.3	94.8 ± 14.1	102.7 ± 27.2	< 0.001
1 hour	725	195.1 ± 28.7	189.2 ± 20.8	195.8 ± 28.2	211.2 ± 43.8	< 0.001
2 hour	723	178.2 ± 29.9	174.9 ± 21.7	178.8 ± 31.0	186.4 ± 43.8	0.009
3 hour	712	129.0 ± 36.5	129.4 ± 31.7	129.2 ± 38.7	126.8 ± 41.5	0.85
GDM diagnosis (weeks)	814	27.3 ± 4.9	28.7 ± 3.6	27.2 ± 4.6	23.6 ± 6.9	< 0.001
Mean glucose across gestation
Fasting	688	88.5 ± 10.3	83.6 ± 8.2	89.4 ± 9.1	95.9 ± 13.9	<0.001
Post breakfast	690	124.5 ± 17.3	115.9 ± 12.5	127.2 ± 16.6	133.4 ± 21.7	<0.001
Post lunch	690	120.7 ± 13.9	114.8 ± 12.2	122.2 ± 12.8	128.1 ± 16.9	<0.001
Post dinner	691	124.0 ± 16.5	115.7 ± 11.8	126.6 ± 16.0	132.7 ± 20.4	<0.001

All data presented as n (%) or mean ± standard deviation

Abbreviations: BMI, body mass index; CHTN, chronic hypertension; GCT, glucose challenge test; MNT, medical nutritional therapy; OGTT, oral glucose tolerance test; GDM, gestational diabetes mellitus.

With regard to birth outcomes, there were proportionately more large for gestational age (LGA) infants among women receiving either glyburide or insulin at delivery (Table 2). Cesarean delivery occurred more frequently in women treated with either glyburide or insulin (Table 2). Composite neonatal morbidity was similar among treatment groups, but the rate of neonatal hypoglycemia was highest among offspring of women treated with glyburide (Table 2). Women who were treated only with MNT were more likely to report breastfeeding at any time point when compared with women treated with either glyburide or insulin (122/293 [41.6%] vs 135/423 [31.9%], vs 34/102 [33.3%]; P = 0.03).

Table 2.

Birth Outcomes, Overall and by Treatment Group

	N if < 816	Total (N = 816)	MNT (n = 293)	Glyburide (n = 421)	Insulin (n = 102)	P value
Gestational age at delivery		38.4 ± 1.8	38.6 ± 2.1	38.3 ± 1.6	37.8 ± 1.7	<0.001
Preterm birth (<37 weeks)	815	116 (14.2%)	37 (12.7%)	56 (13.3%)	23 (22.5%)	0.04
Birth weight (kg)		3.3 ± 0.6	3.2 ± 0.6	3.3 ± 0.6	3.3 ± 0.5	0.04
Placenta weight (grams)	715	467.3 ± 112.4	464.4 ± 119.1	468.9 ± 106.0	468.1 ± 119.9	0.88
Birth weight category
AGA		670 (82.1%)	237 (80.9%)	351 (83.4%)	82 (80.4%)	0.02
SGA		80 (9.8%)	40 (13.7%)	32 (7.6%)	8 (7.8%)
LGA		66 (8.1%)	16 (5.5%)	38 (9.0%)	12 (11.8%)
BMI (kg/m2) at birth	767	12.7 ± 1.5	12.5 ± 1.5	12.8 ± 1.5	12.9 ± 1.2	0.02
Birth weight z-score		0.1 ± 1.0	-0.1 ± 1.0	0.2 ± 1.0	0.4 ± 1.1	<0.001
Macrosomia		53 (6.5%)	15 (5.1%)	29 (6.9%)	9 (8.8%)	0.38
Hypertensive disorders of pregnancy		122 (15.0%)	42 (14.3%)	60 (14.3%)	20 (19.6%)	0.37
Cesarean delivery		317 (38.8%)	93 (31.7%)	175 (41.6%)	49 (48.0%)	0.004
Composite neonatal morbidity		169 (20.7%)	52 (17.7%)	96 (22.8%)	21 (20.6%)	0.26
NICU admission		113 (13.9%)	37 (12.8%)	61 (14.5%)	15 (14.7%)	0.78
Hypoglycemia	812	99 (12.2%)	26 (9.0%)	63 (15.0%)	10 (9.8%)	0.04
RDS		35 (4.3%)	15 (5.2%)	14 (3.3%)	6 (5.9%)	0.5
Hyperbilirubinemia		66 (8.1%)	22 (7.6%)	35 (8.3%)	9 (8.8%)	0.90
Congenital anomaly	811	19 (2.3%)	9 (3.1%)	5 (1.2%)	5 (4.9%)	0.05

All data presented as n (%) or mean ± standard deviation.

Abbreviations: AGA, appropriate for gestational age; BMI, body mass index; LGA, large for gestational age); MNT, medical nutritional therapy; NICU, neonatal intensive care unit; RDS, respiratory distress syndrome; SGA, small for gestational age.

At delivery, birthweight (3.23 ± 0.6 vs 3.3 ± 0.6 vs 3.3 ± 0.5 kg; P = 0.04), birth weight z-score (−0.1 ± 1.0 vs 0.2 ± 1.0 vs 0.4 ± 1.1; P < 0.001), and BMI (12.5 ± 1.5 vs 12.8 ± 1.5 vs 12.9 ± 1.2 kg/m²; P = 0.02) were all higher in offspring of women who were receiving either glyburide or insulin at delivery when compared with women treated with MNT (Table 3). Despite the initial differences at birth, by 6 months of age infant weight, weight z-scores, and BMI were similar among the offspring of women treated with MNT, glyburide, or insulin. These similarities persisted at 1, 2, and 3 years of age (Fig. 2, Table 3). These findings were unchanged when we conducted sensitivity analyses excluding the 27 women who transitioned from glyburide to insulin, and when analyses were conducted based on the initial treatment instead of the treatment at delivery (data not shown).

Table 3.

Measures of Growth at Birth Through 3 Years of Age

	N if < 816	Total	MNT	Glyburide	Insulin	P value
		(N = 816)	(n = 293)	(n = 421)	(n = 102)
Weight (kg)
Birth		3.3 ± 0.6	3.2 ± 0.6	3.3 ± 0.6	3.3 ± 0.5	0.04
6 months	519	7.8 ± 1.1	7.8 ± 1.0	7.8 ± 1.1	7.7 ± 1.0	0.79
1 year	627	9.9 ± 1.4	9.8 ± 1.3	9.9 ± 1.3	10.0 ± 1.6	0.65
2 years	555	12.6 ± 1.7	12.5 ± 1.7	12.7 ± 1.8	12.6 ± 1.8	0.52
3 years	552	15.0 ± 2.2	14.7 ± 1.8	15.0 ± 2.2	15.4 ± 3.2	0.07
Weight z-score
Birth		0.1 ± 1.0	-0.1 ± 1.0	0.2 ± 1.0	0.4 ± 1.1	< 0.001
6 months	518	-0.2 ± 1.1	-0.2 ± 1.0	-0.2 ± 1.1	-0.3 ± 1.1	0.90
1 year	626	-0.2 ± 1.2	-0.3 ± 1.2	-0.2 ± 1.1	-0.2 ± 1.4	0.76
2 years	554	-0.1 ± 1.1	-0.1 ± 1.1	-0.1 ± 1.1	-0.1 ± 1.1	0.79
3 years	542	0.4 ± 1.3	0.4 ± 1.1	0.4 ± 1.3	0.7 ± 1.8	0.18
BMI
Birth	767	12.7 ± 1.5	12.5 ± 1.5	12.8 ± 1.5	12.9 ± 1.2	0.02
6 months	516	17.1 ± 2.5	17.0 ± 1.7	17.2 ± 3.1	17.1 ± 1.4	0.77
1 year	612	17.1 ± 1.7	17.1 ± 1.7	17.1 ± 1.7	17.2 ± 1.9	0.91
2 years	539	16.5 ± 1.8	16.5 ± 1.6	16.6 ± 1.9	16.5 ± 1.8	0.72
3 years	520	16.4 ± 2.1	16.3 ± 2.5	16.3 ± 1.7	16.9 ± 2.4	0.12

Data presented as mean ± standard deviation

Abbreviations: BMI, body mass index; MNT, medical nutritional therapy.

Figure 2.

Childhood weight, weight Z-scores, and BMI over time by GDM treatment types. Figures represent the change in weight (Panel A), weight z-score (panel B) and BMI (panel C) from birth to 3 years of age.

We next evaluated the change over time among women treated with MNT, glyburide, and insulin after adjusting for baseline covariates that differed among treatment groups (mother’s age, pre-pregnancy BMI, first-trimester systolic blood pressure, 50-gram GCT, number of weeks at time of GDM diagnosis, and nulliparity). As expected, weight, weight z-score, and BMI all increased over time (Table 4). When we compared the difference among groups over 3 years there were differences in the weight z-score over time (P = 0.01) by treatment group, but no differences in weight (P = 0.06) or the change in BMI (P = 0.28). Adjusted analyses demonstrated that weight increased faster in the offspring of women treated with insulin compared with MNT (P = 0.02), whereas weight z-score decreased in the offspring of women treated with glyburide compared with MNT (P = 0.05). Because of the nonlinear trajectories of the outcomes, we next conducted analyses that were partitioned at 6 months of age. With regards to weight, there were no differences between treatment groups in the first 6 months, but infants in the insulin group gained more than those in the MNT group (P = 0.01) from 6 months to 3 years. Weight z-scores demonstrated a greater decrease in the first 6 months in the insulin and glyburide groups compared with the MNT group, but there were no statistically significant differences from 6 months to 3 years. The results of these analyses were unchanged when breastfeeding was added to the models (data not shown). Because early childhood overweight/obesity may be an important predictor of obesity in adolescence and adulthood, we also compared the prevalence of BMI > 85th percentile among groups and found no differences at either age 2 (15.7 vs 18.6 vs 15.9%; P = 0.68) or age 3 (18.0 vs 22.1 vs 26.4%; P = 0.31) among the offspring of women treated with MNT, glyburide, or insulin, respectively. There were also no significant differences when we compared BMI > 95th percentile at either age 2 or 3 (data not shown).

Table 4.

Growth Trajectories by Treatment Group Among Women Treated With Medical Nutritional Therapy, Glyburide, and Insulin

	MNT	Glyburide	Insulin		Glyburide vs MNT	Insulin vs MNT	Glyburide vs insulin
	(n = 254)	(n = 374)	(n = 85)
	Mean (95% CI)			Overall P value group*time	Mean (95% CI)
Weight (kg)
0 to 3-year change; P value	11.52 (11.22, 11.83) <0.001	11.74 (11.49, 11.98) <0.001	12.21 (11.70, 12.72)  < 0.001	0.06	0.21 (−0.18, 0.60) 0.28	0.69 (0.10, 1.28) 0.02	0.47 (−0.09, 1.04) 0.10
0 to 6-month change; P value	4.60 (4.46, 4.73)  < 0.001	4.49 (4.37, 4.60)  < 0.001	4.58 (4.34, 4.82)  < 0.001		−0.11 (−0.29, 0.07) 0.23	−0.01 (−0.29, 0.27) 0.93	−0.10 (−0.36, 0.17) 0.48
6-month to 3-year change; P value	6.93 (6.65, 7.20)  < 0.001	7.25 (7.03, 7.47)  < 0.001	7.63 (7.17, 8.09)  < 0.001		0.32 (−0.03, 0.68) 0.08	0.70 (0.16, 1.24) 0.01	−0.38 (−0.89, 0.13) 0.15
Weight z-score
0 to 3-year change; P value	0.48 (0.29, 0.66) <0.001	0.24 (0.09, 0.39) 0.002	0.33 (0.02, 0.64) 0.04	0.01	−0.24 (−0.47, 0.003) 0.05	−0.14 (−0.51, 0.22) 0.43	0.09 (−0.25, 0.43) 0.61
0 to 6-month change; P value	−0.10 (−0.26, 0.06) 0.20	−0.45 (−0.58, −0.32)  < 0.001	−0.53 (−0.81, −0.25)  < 0.001		−0.35 (−0.56, −0.14) 0.001	−0.43 (−0.75, −0.11) 0.01	0.08 (−0.23, 0.38) 0.61
6-month to 3-year change; P value	0.58 (0.40, 0.76) <0.001	0.69 (0.55, 0.84)  < 0.001	0.86 (0.56, 1.17)  < 0.001		0.11 (−0.12, 0.35) 0.34	0.28 (−0.07, 0.64) 0.12	−0.17 (−0.51, 0.17) 0.33
BMI (kg/m 2 )
0 to 3-year change; P value	3.84 (3.50, 4.18) <0.001	3.68 (3.41, 3.96)  < 0.001	4.13 (3.57, 4.70)  < 0.001	0.28	−0.16 (−0.60, 0.28) 0.47	0.29 (−0.37, 0.95) 0.39	0.45 (−0.18, 1.08) 0.16
0 to 6-month change; P value	4.65 (4.34, 4.96)  < 0.001	4.35 (4.09, 4.60)  < 0.001	4.55 (4.02, 5.08) <0.001		−0.30 (−0.70, 0.10) 0.14	−0.10 (−0.71, 0.52) 0.76	−0.21 (−0.79, 0.38) 0.49
6-month to 3-year change; P value	−0.81 (−1.12, −0.50) <0.001	−0.67 (−0.92, −0.42) <0.001	−0.42 (−0.93, 0.09) 0.11		0.14 (−0.25, 0.54) 0.48	0.39 (−0.21, 0.99) 0.20	−0.25 (−0.82, 0.32) 0.40

Linear mixed model with offspring weight as a function of time, treatment group, the interaction of time and treatment group, and the covariates significant in Table 2 (mother’s age, pre-pregnancy BMI, first-trimester systolic blood pressure, 50 g GCT, number of weeks at time of GDM diagnosis, and nulliparity).

Abbreviations: BMI, body mass index; MNT, medical nutritional therapy.

Discussion

We did not observe meaningful differences in infant and early childhood growth outcomes from ages 6 months to 3 years in a population-based cohort of children born to women with GDM treated with either glyburide or insulin when compared to those treated with MNT alone.

These findings are significant because there is a paucity of data regarding long-term childhood outcomes among children exposed to various GDM treatments. While the majority of obesity in adolescents and adults occurs in individuals with normal weight at birth, the outcome of early childhood growth remains clinically relevant, because a substantial component of adolescent obesity is established before 5 years of age. Children who are overweight in kindergarten are 4 times as likely to become obese by eighth grade as compared to those with normal weight (27). In addition, among children who were obese at ages 3 to 5 years, 24% were obese in young adulthood and this increased to 62% if at least 1 parent was obese (28). There were differences in immediate birth outcomes, including higher mean birth weights and risk for LGA birth weight among the offspring of women who were treated with either glyburide or insulin. In addition, women treated with glyburide or insulin had higher rates of cesarean delivery. While overall rates of neonatal morbidity were similar among the offspring of women treated with MNT, glyburide, or insulin, we detected higher rates of neonatal hypoglycemia among women treated with glyburide. It is possible that the higher rates of neonatal hypoglycemia were related to direct effects of glyburide, but data from randomized clinical trials (10, 15) and cohort studies (13, 29) are conflicting. Glyburide is known to cross the placenta, and recent data also suggests that glyburide may increase placental Glucose Transporter 1 (GLUT1) expression, potentially increasing glucose delivery to the fetus (30).

The prevalence of gestational diabetes in increasing, with recent data suggesting that 7% to 9% of pregnant women in the US are affected (31, 32). There are nearly 4 000 000 births yearly in the United States (33), and GDM places a substantial burden on the health care system. There are concerns about rising costs and affordability of insulin (34), which continues to fuel interest in easier-to-use and less-expensive oral agents such as glyburide or metformin. There is a paucity of randomized clinical trials directly comparing glyburide to metformin, and all previous studies were limited by small sample size (35-37). Balsells et al, in a meta-analysis, found that metformin was associated with less maternal weight gain, lower birthweight, less macrosomia, and less LGA when compared to glyburide therapy (12). Another meta-analysis that used a network approach to their meta-analysis and results suggested that metformin has the highest probability of being the most effective treatment when compared with insulin or glyburide (38). However, recent studies of offspring exposed to metformin in utero have produced conflicting results, with 2 of the 3 studies published in children aged 4 to 9 years, whose mothers had either GDM or PCOS suggesting risk for long-term metabolic dysfunction (39-41). The timing of follow-up may also be important, as children who had been exposed to metformin did not differ from children exposed to insulin with regard to weight, height, or abdominal fat at age 2 years (42). These data highlight the importance of additional follow-up studies regarding glyburide use, because both the immediate pregnancy outcomes and long-term growth and metabolic function are important to optimize treatment choices for women with GDM.

This study provides a unique opportunity to examine long-term outcomes among children exposed to glyburide in utero and our findings suggest a potential for the long-term safety profile of glyburide use in pregnancy. Strengths of the study also include detailed data regarding glycemic control in pregnancy and perinatal outcomes. In addition, prior studies of childhood follow-up did not include a group of women exposed to diet alone during pregnancy (39, 42), and we were therefore able to compare the effects of the medication to only dietary modification. However, our study is not without limitations. We lacked a group of women without GDM to establish the pattern of childhood growth in their offspring. Also, only a subset of offspring had follow-up data, and although those with follow-up appeared to be similar to those without it is possible that some bias was introduced. In addition, we were limited to follow-up at age 3 years, and differences in child growth may not appear until late childhood (43). We were not able to perform more detailed metabolic assessment including markers of childhood adiposity or glucose metabolism. Given the association between neonatal hypoglycemia and adverse neurodevelopmental outcomes (44), prospective studies are needed to assess the relationship between GDM treatment and pediatric neurodevelopmental outcomes. It is also possible that there may be residual confounding related to disease severity owing to the observational nature of this study. However, our regression modeling demonstrated that our results were similar even after adjusting for multiple covariates. One factor that may limit generalizability of our results is that the proportion of women in our study who were treated with either glyburide or insulin is higher than the proportion of women requiring treatment in previously published studies (4). However, prior work included only women with mild GDM, and there is also a paucity of data examining treatment patterns when oral agents are used.

Conclusion

Despite fetal growth differences detected at birth, we observed no meaningful differences in childhood growth from 6 months to 3 years among treatment groups, including in the offspring of women treated with glyburide. This will provide clinicians and patients with important information concerning long-term outcomes when evaluating options for GDM treatment. However, further studies are needed to determine if these findings persist beyond 3 years of age and to explore differences in body composition.

Glossary

Abbreviations

BMI

body mass index

GCT

glucose challenge test

GDM

gestational diabetes mellitus

IQR

interquartile range

LGA

large for gestational age

MFM

Maternal Fetal Medicine

MNT

medical nutritional therapy

NICU

neonatal intensive care unit

OGTT

oral glucose tolerance test

Additional Information

Disclosure Summary: The authors have no conflicts of interest to disclose.

Data Availability

Restrictions apply to the availability of some or all data generated or analyzed during this study to preserve patient confidentiality. The corresponding author will on request detail the restrictions and any conditions under which access to some data may be provided.