Pregnancy and diabetes: Emerging insights from contemporary diabetes research

Table 1 Key findings of late-breaking studies on diabetes management in pregnancy of the 84^th American Diabetes Association Scientific Sessions

Study category	Study number	Study design and population	Sample size (n)	Primary outcome (s)	Key methods	Key findings	Clinical and research implications
Basic science/translational studies	1962-LB	Prospective cohort; children (7-11 years) exposed to GDM vs unexposed	204 (110 GDM exposed, 94 unexposed)	Adiposity and brain volumes	Brain MRI, adiposity measurements over 6 years, mixed-effects models	In utero GDM exposure was associated with increased adiposity (BMI, body fat percentage, waist circumference) and increased growth of total cortex and gray matter in offspring	Provides a mechanistic link between GDM exposure and adverse metabolic and neural outcomes in children. Population/setting: United States multiethnic cohort; IADPSG criteria. Risk-of-Bias Score: NOS = 7/9 (moderate)
	1965-LB	Prospective; mother-baby dyads (BMI < 25, obese, GDM)	39 dyads (13 control, 14 obese, 12 GDM)	Placental transcriptomic changes, cord blood metabolic profiles	RNA sequencing (placenta), metabolic profiling (cord blood)	GDM and obesity were associated with upregulation of genes involved in nutrient transport, inflammation, and methylation, as well as increased INSR expression	Suggests methylation and epigenetic changes may contribute to intergenerational transmission of metabolic risks. Population/setting: Japan (animal model); GLP-1 treatment. Risk-of-Bias Score: ARRIVE 2.0 = moderate
	260-OR	Observational; 3rd trimester biopsies (NT, GDM, PE)	30 (GDM), 7 (PE), 30 (NT)	Adipocyte size, AT fibrosis, mRNA expression, insulin signaling	Adipocyte size measurement, AT fibrosis assessment, mRNA expression analysis, insulin signaling (immunoblotting)	GDM was associated with increased visceral adipose tissue diameter, hypertrophy, and HIF1A mRNA expression. Both GDM and preeclampsia showed decreased insulin-stimulated Akt phosphorylation in subcutaneous adipose tissue. Preeclampsia did not affect visceral adipose tissue	Indicates that GDM and preeclampsia have distinct effects on adipose tissue signaling and function. Population/setting: Obese United States cohort; WHO 2013 GDM criteria. Risk-of-Bias Score: NOS = 6/9 (moderate)
	200-OR	Prospective cohort; women with/without GDM postpartum	20 (10 GDM, 10 control)	Islet-cell function (glucose, insulin, C-peptide, glucagon)	75 g OGTT at 3, 6, 12 months postpartum	Women with prior GDM showed increased fasting glucose, insulin, C-peptide, and glucagon at 3 months postpartum, but no differences were observed by 6 months	Early postpartum islet-cell dysfunction may increase long-term metabolic risk in women with a history of GDM. Population/setting: United Kingdom cohort; postpartum GDM analysis. Risk-of-Bias Score: NOS = 7/9 (moderate)
	199-OR	Interventional (mice); β-cell IRS1-knockout mice	Animal model	β-cell compensation, glucose levels	Islet RNA transcriptome analysis	IRS1 deficiency disrupted the IRS1-GATA4-Reg1/Reg3a pathway, impaired β-cell compensation, and led to GDM	Understanding this pathway may lead to new prevention and treatment strategies for GDM. Population/setting: Mouse model; IRS1 KO strain. Risk-of-Bias Score: ARRIVE 2.0 = low
Clinical/epidemiological studies	1969-LB	Prospective cohort; South Asian women postpartum	49	Insulin resistance and deficiency	75 g OGTT, insulin levels at 6 weeks and 6 months postpartum, HOMA-IR, IGI, oDI	Postpartum glucose intolerance was associated with insulin deficiency (decreased IGI, oDI) but not with changes in insulin resistance (HOMA-IR)	Postpartum insulin deficiency may be a better predictor of type 2 diabetes risk in South Asian women than insulin resistance. Population/setting: South Asian cohort; Chennai; postpartum glucose. Risk-of-Bias Score: NOS = 8/9 (low)
	62-OR	Retrospective cohort; T1DM pregnancies	86	Adverse pregnancy outcomes	CGM metrics (TIR, TAR, TBR, CV) from 28-39 weeks, linear mixed-effects model	HDP were associated with increased TAR and decreased TIR at 28 weeks	Early 3rd trimester CGM metrics are critical for preventing adverse pregnancy outcomes in T1DM pregnancies. Population/setting: United States T1DM pregnancies; Dexcom CGM. Risk-of-Bias Score: NOS = 7/9 (moderate)
	63-OR	Prospective observational; pregnant women	760	GDM, LGA, HDP	CGM data, logistic and elastic net regression	CGM-based models (TA140) predicted GDM, LGA, and HDP	CGM at 13-14 weeks could potentially serve as an alternative to the OGTT at 24-28 weeks for risk prediction. Population/setting: Multisite CGM trial; 760 pregnant women. Risk-of-Bias Score: NOS = 8/9 (low)
	64-OR	Prospective observational; pregnant women with glucose intolerance	Varies by group	TIR	Single-blinded CGM, GCT, OGTT	Abnormal GCT or OGTT results were associated with decreased TIR on CGM	Abnormal GCT/OGTT values should not be discounted, even if isolated findings. Population/setting: CGM data; abnormal GCT/OGTT cases. Risk-of-Bias Score: NOS = 7/9 (moderate)
	65-OR	Interventional; women with GDM (BMI > 25)	432	Adverse pregnancy outcomes	CGM, logistic regression	CGM metrics (mean glucose, TIR, TAR) at 29 weeks predicted adverse pregnancy outcomes	CGM metrics can predict adverse pregnancy outcomes in women with GDM. Population/setting: CGM metrics in overweight women with GDM. Risk-of-Bias Score: NOS = 7/9 (moderate)
	66-OR	Prospective observational; GDM mothers and neonates	13 mother-neonate dyads	Neonatal hypoglycemia	Masked CGM during labor and delivery, neonatal thigh CGM	CGM identified neonatal hypoglycemia that was missed by routine practices	CGM is potentially useful for identifying neonatal hypoglycemia. Population/setting: Neonatal CGM; 13 mother-infant dyads. Risk-of-Bias Score: NOS = 6/9 (moderate)
	67-OR	Prospective observational; uncomplicated pregnancies	157	Postprandial glucose patterns	Blinded CGM, smartphone app	HDP were associated with increased postprandial glucose excursions	Suggests an association between HDP and elevated postprandial glucose excursions. Population/setting: Uncomplicated pregnancies; CGM data. Risk-of-Bias Score: NOS = 8/9 (low)
	1976-LB	Prospective cohort; women with GDM history	2999	T2D risk	Cox regression, metabolic biomarkers	Shorter sleep duration (≤ 6 hours) and frequent snoring were associated with increased T2D risk	Sleep characteristics are important predictors of T2D risk in women with a history of GDM. Population/setting: NHS II cohort; sleep and T2D risk. Risk-of-Bias Score: NOS = 8/9 (low)
	1978-LB	Multiracial cohort; pregnant women	321	Cardiometabolic biomarkers	Sleep duration assessment, cardiometabolic biomarker measurements, linear mixed models	Shorter sleep duration was associated with increased glucose and HbA1c. Longer sleep duration was associated with increased insulin at the first visit. Shorter sleep duration at the second visit was associated with increased glucose, and longer sleep duration at the second visit was associated with decreased HbA1c	Adequate sleep is crucial for maintaining cardiometabolic health during pregnancy. Population/setting: United States NICHD cohort; sleep/cardiometabolic markers. Risk-of-Bias Score: NOS = 7/9 (moderate)
	202-OR	Prospective cohort; pregnant women without early GDM	2685	LGDM	HbA1c, OGTT, ROC assessment	1HBG level best predicted LGDM, but OGTT was still needed.	1HBG is a good predictor of LGDM, but OGTT remains essential for diagnosis. Population/setting: Prospective OGTT cohort; risk factors for LGDM. Risk-of-Bias Score: NOS = 8/9 (low)
	203-OR	Randomized controlled trial; pregnant women with 1 abnormal OGTT value	827	Metabolic and clinical outcomes	Propensity score models with IPTW	Women with 1 abnormal OGTT value had an increased risk of LGA compared to women with normal glucose tolerance and those with GDM	Women with 1 abnormal OGTT value have a metabolic profile closer to GDM and a higher risk of LGA. Population/setting: RCT; Carpenter-Coustan 1 abnormal OGTT. Risk-of-Bias Score: RoB-2 = moderate
	1975-LB	Randomized trial; GDM patients	107	Glycemic control	RT-CGM vs SMBG	RT-CGM was associated with decreased mean glucose and TBR54 at 36 weeks, but increased medication use	RT-CGM may offer benefits in lowering mean glucose and TBR54 in GDM, although it may be associated with increased medication use. Population/setting: RCT; CGM vs SMBG in GDM patients. Risk-of-Bias Score: RoB-2 = low
	259-OR	Randomized trial; pregnant women with GDM	111	TIR	Real-time CGM vs SMBG, two-sample t-tests	Real-time CGM was associated with higher TIR than SMBG, particularly during daytime	CGM is superior to SMBG in maintaining glucose range in women with GDM. Population/setting: RCT; CGM vs SMBG; TIR analysis. Risk-of-Bias Score: RoB-2 = low
	198-OR	Prospective cohort; pregnant women undergoing OGTT	1308	GDM diagnosis	Enhanced vs standard glucose processing	Enhanced glucose processing was associated with increased GDM diagnosis rates	Accurate glucose processing is crucial for the appropriate diagnosis of GDM. Population/setting: United Kingdom cohort; HbA1c vs OGTT. Risk-of-Bias Score: NOS = 7/9 (moderate)
Therapeutic interventions and outcomes	1973-LB	Prospective cohort; postpartum women with GDM	50	Postpartum dysglycemia	Postpartum CGM vs OGTT	Postpartum CGM was feasible and acceptable, and the percentage of time > 180 mg/dL predicted OGTT dysglycemia	CGM is a useful postpartum screening tool for dysglycemia. Population/setting: Prospective cohort; GGI and preeclampsia. Risk-of-Bias Score: NOS = 7/9 (moderate)
	1963-LB	Interventional (mice); diabetic mothers	Animal model	Cognitive deficits in offspring	GLP-1 agonist treatment, placental development assessment, behavioral testing of offspring, immunohistochemistry, single-cell RNA sequencing, qRT-PCR, in vitro studies	Prenatal GLP-1 agonist treatment mitigated cognitive deficits in offspring of diabetic mothers	GLP-1 agonists may have a role in preventing cognitive deficits in offspring of mothers with diabetes. Population/setting: HbA1c analysis in United States cohort. Risk-of-Bias Score: NOS = 8/9 (low)
	1964-LB	Prospective cohort; women with glucose intolerance	106	Preeclampsia development	Fasting lactate, insulin, glucose, HOMA-IR measurements	Increased fasting venous lactate was predictive of preeclampsia development in women with glucose intolerance	Fasting lactate may be a useful predictor of preeclampsia risk in women with glucose intolerance. Population/setting: T1DM women; CGM-HbA1c profiles. Risk-of-Bias Score: NOS = 8/9 (low)
Maternal and neonatal health	1970-LB	Prospective cohort; pregnant women at delivery	609	Adverse obstetric and perinatal outcomes	Third-trimester HbA1c measurement	HbA1c > 5.8% was associated with increased adverse obstetric and perinatal outcomes (C-section, hemorrhage, macrosomia, NICU admission, etc.)	HbA1c > 5.8% in the third trimester is a risk factor for pregnancy complications. Population/setting: Retrospective United States cohort; overt DM timing. Risk-of-Bias Score: NOS = 7/9 (moderate)
	1971-LB	Observational; T1DM pregnancies	112	Fasting and postprandial glucose contributions to hyperglycemia	CGM data analysis	Fasting hyperglycemia is a major contributor to overall hyperglycemia in T1DM pregnancies	Optimizing insulin regimens to reduce fasting hyperglycemia may improve outcomes in T1DM pregnancies. Population/setting: EHR + ML models for GDM; United States data. Risk-of-Bias Score: NOS = 8/9 (low)
	1972-LB	Retrospective cohort; women with overt diabetes	646	Pregnancy outcomes	Comparison of outcomes based on timing of diagnosis	Early diagnosis of overt diabetes was associated with better metabolic control (decreased weight gain, decreased 3rd-trimester HbA1c) but similar overall pregnancy outcomes compared to later diagnosis	Early screening for hyperglycemia in pregnancy is important for achieving better metabolic control. Population/setting: Louisiana cohort; Medicaid + COVID impact. Risk-of-Bias Score: NOS = 7/9 (moderate)
	1968-LB	Retrospective; pregnant women with GDM	About 27500	GDM prediction	EHR data, machine learning models (logistic regression, etc.)	Logistic regression models using EHR data can predict GDM in the first trimester	EHRs and machine learning can aid in early GDM prediction. Population/setting: RCT; energy diet in GDM (United Kingdom DiGest). Risk-of-Bias Score: RoB-2 = low
	1974-LB	Retrospective cohort; pregnant women in Louisiana	110447	Adverse pregnancy outcomes	Medicaid claims data, logistic regression	Maternal hyperglycemia and hypertensive disorders during the COVID-19 pandemic were associated with increased adverse pregnancy outcomes (LGA, macrosomia)	Tailored strategies are needed for managing high-risk pregnancies during crises such as the COVID-19 pandemic. Population/setting: Early GDM management; capillary glucose data. Risk-of-Bias Score: NOS = 8/9 (low)
	257-OR	Randomized controlled trial; women with GDM (BMI ≥ 25)	423	Maternal weight change, offspring birth weight	Dietary intervention (standard vs reduced-energy diet)	A reduced-energy diet was associated with decreased insulin needs, but had no significant effect on maternal weight change or offspring birth weight compared to a standard diet	Standard and reduced-energy diets have similar effects on maternal weight and offspring birth weight in women with GDM. Population/setting: Case study; congenital lipodystrophy, AGPAT2. Risk-of-Bias Score: ARRIVE 2.0 = moderate
	258-OR	Prospective cohort; women with early GDM	399	Glycemic control, pregnancy complications	Capillary blood glucose monitoring, comparison of early vs late treatment	Early treatment of early GDM was associated with better glycemic control (decreased mean glucose, increased optimal glycemia) and fewer pregnancy complications compared to late treatment	Early diagnosis and treatment of GDM are crucial for optimizing glycemic control and reducing complications. Population/setting: Obese pregnancy cohort; triglycerides, CGM. Risk-of-Bias Score: NOS = 7/9 (moderate)
	1977-LB	Case study; woman with CGL	1	Pregnancy outcomes	Clinical observation of two pregnancies	Successful pregnancies are possible in women with CGL without leptin therapy	CGL can present with varying degrees of metabolic severity, and successful pregnancy outcomes are possible. Population/setting: MRI + adiposity; LA-based United States cohort. Risk-of-Bias Score: NOS = 7/9 (moderate)
	201-OR	Observational cohort; overweight/obese pregnant women	31	Correlation between early pregnancy triglycerides and glucose metrics at 28 weeks	FTG and PPTG measurement, CGM at 28 weeks	Fasting and postprandial triglycerides in early pregnancy correlated with mean glucose and TIR at 28 weeks	Triglyceride levels in early pregnancy may help identify women at risk for later hyperglycemia. Population/setting: Placenta transcriptomics; obesity/GDM cohort. Risk-of-Bias Score: NOS = 7/9 (moderate)

GDM: Gestational diabetes mellitus; BMI: Body mass index; MRI: Magnetic resonance imaging; AT: Adipose tissue; SAT: Subcutaneous adipose tissue; VAT: Visceral adipose tissue; HIF1A: Hypoxia-inducible factor 1-alpha; OGTT: Oral glucose tolerance test; IGI: Insulinogenic index; oDI: Oral disposition index; HOMA-IR: Homeostatic model assessment of insulin resistance; T1DM: Type 1 diabetes mellitus; CGM: Continuous glucose monitoring; TIR: Time in range; TAR: Time above range; TBR/TBR54: Time below range (< 54 mg/dL); CV: Coefficient of variation; TA140: Time above 140 mg/dL; GCT: Glucose challenge test; LGA: Large for gestational age; HDP: Hypertensive disorders of pregnancy; PE: Preeclampsia; LGDM: Late gestational diabetes mellitus; 1HBG: One-hour blood glucose; SMBG: Self-monitored blood glucose; RT-CGM: Real-time continuous glucose monitoring; NICU: Neonatal intensive care unit; EHR: Electronic health records; FTG: Fasting triglycerides; PPTG: Postprandial triglycerides; GLP-1: Glucagon-like peptide 1; COVID-19: Coronavirus disease 2019.