Fetal sex and the development of gestational diabetes mellitus in gravidae with multiple gestation pregnancies

Abstract

Introduction

There is an increasing incidence of pregnancies with twin gestations. One outcome more likely to occur with multiple gestations is gestational diabetes mellitus. Studies have suggested that in singleton pregnancies, fetal sex may affect insulin resistance. However, the effects of fetal sex in twins and the development of gestational diabetes mellitus are unknown. We hypothesized that rates of gestational diabetes mellitus and degree of insulin resistance might vary in twin gestations based on the fetal sex pairing: male–male, male–female or female–female. We aimed to employ a large population‐based database to ascertain any correlations between fetal sex and gestational diabetes mellitus in multifetal gestations.

Material and methods

A two‐hospital, single academic institution database comprised of over 39 000 participants with pregnancy data from August 2011 to January 2022 was employed. All twin deliveries of live‐born neonates >24 weeks’ gestational age from gravidae without preexisting diabetes or twin–twin transfusion syndrome were included. Entries were then grouped based on the fetal sex of the pairing. The presence or absence of gestational diabetes and type of gestational diabetes – diet‐controlled (gestational diabetes mellitus classification A1) vs medication‐controlled (gestational diabetes mellitus classification A2) – were identified. Statistical analysis was performed using a generalized linear mixed method, and a P‐value ≤0.05 was considered statistically significant.

Results

We identified 1924 twin deliveries that met the inclusion criteria in our database (male–male =652; male–female = 638; female–female = 634). We found no association between fetal sex pairing and the development of gestational diabetes mellitus. There was a significant association between the fetal sex pairing and the type of gestational diabetes mellitus developed, with 32.0% of male–male twins, 33.3% of male–female twins and 58.3% of the female–female twin deliveries associated with medication‐controlled gestational diabetes classification A2: male–female vs female–female (P = 0.05) and male–male vs female–female (P = 0.046).

Conclusions

While gestational diabetes mellitus is of multifactorial origin, we found a significant association between the fetal sex pairing and the treatment needed for gravidae with twins who develop gestational diabetes mellitus. A higher proportion of female–female twins was associated with gestational diabetes classification A2 compared with male–female or male–male deliveries. Further research on the physiology driving this association is warranted.

While gestational diabetes mellitus is of multifactorial origin, this study was designed to investigate whether gestational diabetes mellitus and degree of insulin resistance might vary in twin gestations based on the fetal sex pairing: male–male, male–female or female–female. There was no relationship between the development of gestational diabetes mellitus and the fetal sex pairing observed in gravidae with twin gestations, with equal proportions of gravidae with male–male, male–female and female–female twins developing the disease. However, in gravidae with twins that develop gestational diabetes mellitus, a higher proportion of deliveries of female–female twins was associated with medication‐controlled gestational diabetes class A2 compared with deliveries of male–female or male–male twins.

Abbreviations

BMI

body mass index (kg/m²)

DCDA

dichorionic‐diamniotic

F‐F

female–female

GDM

gestational diabetes mellitus

GDMA1

gestational diabetes mellitus class A1

GDMA2

gestational diabetes mellitus class A2

GWG

gestational weight gain

MCDA

monochorionic diamniotic

MCMA

monochorionic monoamniotic

M‐F

male–female

MG

multiple gestation

M‐M

male–male

Key message.

While gestational diabetes mellitus is of multifactorial origin, a higher proportion of deliveries of female–female twins required medical management to achieve glycemic control compared with deliveries of male–male or male–female twins complicated by gestational diabetes mellitus.

1. INTRODUCTION

Twin or multiple gestation (MG) pregnancies confer an increased risk of adverse outcomes in pregnancy. Over the last decades, there has been a substantial increase in MGs, partly because of assisted reproductive technologies. ¹ While the twinning rate (the number of twin deliveries divided by the total number of deliveries, per 1000) appears to have peaked in the USA in 2014 with a twinning rate of 17.27, data from the Human Multiple Births Database indicate a substantial USA twinning rate of 15.83 in 2021. ² Pregnant persons (gravidae) with twins are at risk of complications similar to singleton pregnancies, such as preeclampsia, gestational diabetes mellitus (GDM), preterm birth and growth restriction, but often at higher rates. ³ Additionally, the pathophysiologic condition for some complications likely differs in twins from those in singleton gestations. In twins, the doubling of placental mass and subsequently higher levels of placental hormones with diabetogenic effects potentially contribute to the higher rates of GDM observed in MG pregnancies. ⁴ Others have postulated that excess carbohydrates, inadequate protein and other micronutrient deficiencies may contribute to higher rates of prematurity, preeclampsia, growth restriction and possibly GDM when comparing twins with singletons. ⁵ Consequently, when adjusting for confounders, studies have shown a twofold increase in risk of developing GDM in pregnancies complicated by twins compared with singletons. ⁶

However, the effects of GDM on fetal outcomes in twin gestations can differ from those of singleton pregnancies. In singleton pregnancies complicated by GDM, accelerated fetal growth and macrosomia are the main adverse effects associated with metabolic dysfunction. ⁷ Unlike singletons, pregnancies complicated by twins with GDM are less likely to be associated with accelerated fetal growth. ⁷ Thus, it is essential to ascertain the factors driving GDM and subsequent fetal outcomes in twin gestations, both similar and unique to pregnancies with singleton gestations.

Fetal sex is one factor potentially influencing the incidence of obstetrical complications like GDM. A complex fetoplacental hormonal milieu exists that is fetal‐sex dependent. ⁸ Thus, it is plausible that maternal metabolism could be affected in a way that is sexually dimorphic vis‐à‐vis the placenta. As such, previous studies have identified a 5% increase in the risk of GDM associated with gravidae with singleton pregnancies carrying a male fetus. ⁹ This excess risk is thought to be related to a reduction in pancreatic beta‐cell function and an ensuing impairment in maternal glucose metabolism. ¹⁰ Maternal pancreatic beta cell mass and function are also regulated by the endocrine function of the placenta. ¹¹ For this reason, others have postulated that the degree of maternal insulin resistance in singletons is potentially associated with fetal sex. ¹² , ¹³ However, the impact of fetal sex on the development of GDM and insulin sensitivity in MG pregnancies with two fetuses has not been well studied. We therefore employed a large population‐based database to ascertain any correlations between fetal sex and GDM in multifetal twin gestations.

2. MATERIAL AND METHODS

This retrospective cohort study was designed to evaluate the rates of GDM in gravidae with twin gestations based on fetal sex pairing, ie male–male (M‐M), male–female (M‐F) or female–female (F‐F). Additional obstetrical outcomes were examined in our cohort as part of the secondary analysis.

We used a two‐hospital, single academic institution database of deliveries beginning August 2011 in Houston, Texas. PeriBank is a comprehensive institutional database and biobank curated at Baylor College of Medicine that focuses on detailed clinical data and accompanying specimens collected at the time of delivery. A detailed description of Peribank has been published previously. ¹⁴ Prenatal records, electronic medical records and in‐person interviews were utilized to obtain up to 4900 clinical data variables on over 39 000 deliveries. Data quality was determined by regular verification of a subset of inserted clinical data and by at least one board‐certified maternal‐fetal medicine physician‐scientist (KMA), as previously published. ¹⁴

Gravidae included in the current study were enrolled in the Peribank database between August 2011 and January 2022 (n = 39 778 deliveries). Each entry represented the delivery of a live‐born neonate, and only deliveries of live‐born neonates ≥24 weeks’ gestational age (GA) were included. Participants were included if the number of neonates associated with the delivery equaled two, signifying a twin delivery. Additionally, gravidae with a preexisting diabetes diagnosis were excluded. Chorionicity, including monochorionic monoamniotic (MCMA), monochorionic‐diamniotic (MCDA) and dichorionic‐diamniotic (DCDA), was noted. Deliveries with an associated twin–twin transfusion syndrome diagnosed antenatally were excluded. In total, 1924 deliveries met our inclusion criteria (see Figure 1). The deliveries were divided into three groups based on the fetal sex of the twin pairs. Of the 1924 deliveries that met our inclusion criteria, 652 were deliveries of M‐M twins, 638 were deliveries of M‐F twins and 634 were deliveries of F‐F twins.

FIGURE 1.

Inclusion and exclusion criteria utilized.

2.1. Statistical analysis

The primary outcome in this study was the presence or absence of GDM in relation to the primary exposure of fetal sex in our study population of deliveries of twin gestations (M‐M, M‐F and F‐F). All clinical and demographic data were obtained from PeriBank, and participants’ electronic medical records were not retrospectively accessed for this study. Groups of interest were compared by chi‐square or Fisher's exact test for categorical variables and Wilcoxon rank sum test for non‐normal continuous variables. All deliveries used in the analysis indicated the presence or absence of a GDM diagnosis. In the deliveries with an associated GDM diagnosis, our secondary outcome of interest was the associated White classification ¹⁵ of GDM based on fetal sex pairing. Gestational diabetes mellitus class A1 (GDMA1) included GDM diagnoses managed with diet and exercise. Gestational diabetes mellitus class A2 (GDMA2) included GDM diagnoses managed with medication (ie oral agents or insulin therapy). Four deliveries in the M‐M group that developed GDM did not specify the type of GDM and were excluded from the secondary analysis. Statistical analysis was performed using a generalized linear mixed method. P ≤ 0.05 was considered statistically significant. All analyses were performed in SAS 9.4.

2.2. Ethics statement

This study utilized PeriBank, an institutional database that was initially approved by the Baylor College of Medicine Institutional Review Board (H‐26364) on 10 September 2011 and has been continuously renewed and active (most recent re‐approval 8 October 2021). After obtaining written consent, Peribank research personnel enrolled gravidae at the time of admission for delivery.

3. RESULTS

During the study period, 39 778 deliveries were captured with completed data in our PeriBank database. A total of 1924 deliveries of twins were identified as our study sample. The maternal characteristics of the three groups of twins are displayed in Table 1. The proportions of MCMA, MCDA and DCDA twin pregnancies were similar between M‐M and F‐F twin pairings. All M‐F twin pairs were DCDA. Of the 652 M‐M deliveries, 54 (8.3%) developed GDM. Of the 638 M‐F deliveries, 54 (8.5%) developed GDM. Of the 634 F‐F deliveries, 72 (11.4%) developed GDM. There was no association between fetal sex pairing and the development of GDM on univariate analysis when comparing each of the sex‐paired groups: M‐M vs M‐F (P = 0.92); M‐F vs F‐F (P = 0.23) and M‐M vs F‐F (P = 0.19) (Table 2). Additionally, we performed a multivariate analysis adjusted for maternal age, preconception body mass index (BMI), and parity in the regression model. Again, we found no association between fetal sex pairing and the development of GDM: M‐M vs M‐F (P = 0.81), M‐F vs F‐F (P = 0.20); and M‐M vs F‐F (P = 0.30) (Table 2).

TABLE 1.

Maternal characteristics.

	Fetal sex pairing			P‐value
Maternal characteristic	Male–male twins (M‐M) (n = 652)	Male–female twins (M‐F) (n = 638)	Female–female twins (F‐F) (n = 634)
Age
Mean (SE)	30.31 (0.32)	31.29 (0.33)	30.64 (0.33)	M‐M vs M‐F: 0.037 a
Preconception BMI
Median (IQR)	24.8 (21.9–29.3)	25.4 (21.9–30.4)	26.4 (22.8–31.2)	M‐M vs M‐F: 0.006 a
Parity, n (%)
0	248 (38.0%)	254 (39.8%)	244 (38.5%)
1	202 (31.0%)	176 (27.6%)	156 (24.6%)
2 or more	202 (31.0%)	208 (32.6%)	234 (36.9%)
Twin chorionicity, n (%)				M‐M vs M‐F: 0.01 a M‐F vs F‐F: 0.01 a
Monochorionic‐monoamniotic	6 (0.92%)	0	14 (2.2%)
Monochorionic‐diamniotic	230 (35.4%)	0	214 (33.8%)
Dichorionic‐diamniotic	414 (63.7%)	624 (100%)	406 (64.0%)

Note: All groups (M‐M, M‐F and F‐F) were compared with each other (M‐M vs M‐F, M‐F vs F‐F, and M‐M vs F‐F).

^a P‐values ≤ 0.05 considered significant. Comparisons where P ≤ 0.05 are shown; otherwise, P‐values were not statistically significant and are not shown.

Abbreviations: BMI, body mass index (kg/m²); IQR, interquartile range; SE, standard error.

TABLE 2.

Twin gestations and rates of gestational diabetes mellitus (GDM). ^b

Characteristic	Outcome		P univariate	P multivariate a
Fetal sex pairing	GDM diagnosis	No‐GDM diagnosis	M‐M vs M‐F: 0.92 M‐F vs F‐F 0.23 M‐M vs F‐F 0.19	M‐M vs M‐F: 0.81 M‐F vs F‐F: 0.20 M‐M vs F‐F: 0.30
Male–male Twins (n = 652)	54 (8.3%)	598 (91.7%)
Male–female Twins (n = 638)	54 (8.5%)	584 (91.7%)
Female–female Twins (n = 634)	72 (11.4%)	562 (88.6%)

^a P‐value is adjusted for maternal age, preconception BMI and parity in the regression model.

^b Diagnosed by two or more elevated values for the 3‐hour 100‐g oral glucose tolerance test (GTT) based on Carpenter/Coustan values. ¹⁶

For a secondary analysis, we analyzed rates of GDMA1 and GDMA2 based on fetal sex pairing in the deliveries with an associated GDM diagnosis (n = 176). In the M‐M group (n = 50), 34 deliveries (68.0%) had an associated GDMA1 diagnosis and 16 (32.0%) were classified as GDMA2. In the M‐F cohort (n = 54), 36 deliveries (66.7%) were identified as GDMA1 and 18 (33.3%) had an associated GDMA2 diagnosis. In the F‐F cohort (n = 72), 30 deliveries (41.7%) were associated with a GDMA1 diagnosis and 42 (58.3%) were classified as GDMA2. There was a difference in univariate analysis of rates of maternal GDMA1 vs GDMA2 based on fetal sex when comparing F‐F paired twins with M‐M or M‐F paired twins: M‐F vs F‐F (P = 0.05) and M‐M vs F‐F (P = 0.046)] (Table 3). However, this difference was not noted when comparing rates of GDMA1 vs GDMA2 in M‐F twin pairings as compared with M‐M twin pairings: M‐M vs M‐F (P = 0.918)] (Table 3). Additionally, we performed a multivariate analysis adjusted for maternal age, preconception BMI and parity in the regression model to compare rates of GDMA1 and GDMA2. This statistically significant difference in rates of type of GDM was noted again on multivariate analysis when comparing M‐M vs F‐F twin pairings (P = 0.049) and M‐F with F‐F twin pairings (P = 0.042) (Table 3).

TABLE 3.

Twin gestations and White Classification of gestational diabetes mellitus (GDM).

Characteristic	Outcome		P univariate	P multivariate a
Fetal sex pairing	GDMA1	GDMA2	M‐M vs M‐F: 0.918 M‐F vs F‐F: 0.05 b M‐M vs F‐F: 0.046 b	M‐M vs M‐F: 0.750 M‐F vs F‐F: 0.042 b M‐M vs F‐F: 0.049 b
Male–male Twins (n = 50)	34 (68.0%)	16 (32.0%)
Male–female Twins (n = 54)	36 (66.7%)	18 (33.3%)
Female–female Twins (n = 72)	30 (41.7%)	42 (58.3%)

^a P‐value is adjusted for maternal age, preconception BMI and parity in the regression model.

^b P ≤ 0.05 considered significant.

4. DISCUSSION

Our initial analysis demonstrated no association between fetal sex twin pairing and the development of GDM. However, parsing by GDM subgroup diagnosis revealed a difference in rates of GDMA1 vs GDMA2 based on fetal sex pairing that remained when adjusting for maternal age, preconception BMI and parity. We determined that MG of two female fetuses was significantly associated with the development of GDMA2 compared to MG that contained at least one male fetus. Although GDM is of multifactorial origin, a higher proportion of deliveries of F‐F twins required medical management of GDM when compared with M‐M or M‐F twins complicated by GDM.

The clinical effects of fetal sex on metabolic dysfunction, specifically GDM, have not been well‐characterized in singleton or twin gestations. In singletons, Retnakaran et al. reported that gravidae carrying a male fetus showed an increased risk of GDM. ¹⁰ In contrast, others found greater insulin resistance in women carrying male fetuses but no differences in the rates of GDM. ¹⁷ Postpartum, gravidae with female singletons have been associated with higher risks of developing T2DM in the interpregnancy period. ¹⁰ A significant proportion of women, ranging from 3% to 65%, with a previous GDM diagnosis develop type 2 diabetes mellitus (T2DM) within 5–6 years after the pregnancy. ¹⁸ GDM can also adversely affect the long‐term development of the offspring, with studies showing impaired glucose tolerance and obesity in childhood. ¹⁹ , ²⁰ Thus, these studies suggest a possible association between fetal sex and glycemic control with potential short‐ and long‐term effects in singletons. Our understanding of the impact and treatment of insulin resistance – and subsequently any potential fetal sex effect – in twins is even more limited. The Society for Maternal‐Fetal Medicine recently highlighted that additional research and studies are necessary to uniquely examine the diagnosis and treatment of GDM in twin gestations. ³ To our knowledge, this study is the first to examine the relationship between GDM and fetal sex in twin gestations. A better understanding of the role fetal sex potentially plays in the development or severity of GDM in gravidae with twins is necessary to inform shared decision‐making with patients and offer timely interventions with the potential to improve maternal‐fetal outcomes.

There is an increased risk of fetal growth restriction in MG pregnancies compared with singleton pregnancies. ³ Some studies have suggested that a mild increase in serum glucose in gravidae with mild GDM may benefit twins and decrease the risk of fetal growth restriction. Ashwal et al. found that GDMA1 was associated with accelerated asymmetric fetal growth in singleton pregnancies but not in MG pregnancies. ⁷ They hypothesize that GDM in twin gestations may be milder and easier to control with dietary intervention than GDM in singleton gestation because gravidae with twins have a lesser degree of beta‐cell dysfunction than women with singleton gestations complicated by GDM. ⁷ Indeed, in twin gestations complicated by GDM with at least one male fetus, we found a greater proportion of deliveries were associated with GDMA1 rather than GDMA2. Further studies are needed to explore the effect of F‐F pairing on maternal beta‐cell dysfunction and fetal growth.

Clinicians have long been attempting to determine antenatal factors that could be useful in assessing the likelihood of patients with a new GDM diagnosis needing treatment with medical therapy. Identifying factors associated with the need for insulin therapy is of essential importance in planning for and providing obstetrical care. In singletons, maternal age, family history of diabetes, obesity, prior GDM and fasting glucose levels are some factors thought to be associated with an increased need for insulin therapy. ²¹ One of the most consistently reported findings in singletons is the association between pre‐pregnancy obesity and an increased need for medical therapy. ²² Other studies have identified that for every additional 2 kg of excessive gestational weight gain (GWG) during GDM management, there was a 32% greater likelihood of insulin being initiated in singletons after adjustment for cofounders. ²³ However, it is important to note that these studies focus on predicting progression to GDMA2 in singleton pregnancies, not MGs. Looking at pre‐pregnancy biometrics in our study, gravidae with F‐F twins were noted to have the highest median preconception BMI of the three groups of twins and were considered “overweight”, with a median BMI of 26.3 kg/m² (vs 24.8 kg/m² for M‐M and 25.4 kg/m² for M‐F twins). To account for this difference, we performed a multivariate analysis that included an adjustment for preconception BMI. Even with this adjustment, the association between fetal sex and the development of GDMA2 remained. To account for any confounding effects of excessive GWG on the progression to GDMA2 noted in the F‐F twin cohort, we performed a post‐hoc analysis of GWG. GWG associated with each delivery in the M‐M, M‐F and F‐F groups was calculated and compared with the Institute of Medicine (IOM)'s recommendations regarding appropriate GWG for normal weight, overweight and gravidae with obesity with MG. ²⁴ We found similar proportions of patients below or meeting appropriate GWG guidelines in the F‐F group compared with the M‐F or M‐M groups (Table 4). Additionally, there was no difference in the proportion of gravidae exceeding recommended GWG when comparing F‐F with the M‐F or M‐M groups: M‐F vs F‐F (P = 0.142) and M‐M vs F‐F (P = 0.43 9) (Table 4).

TABLE 4.

Twin gestations and recommendations for gestational weight gain (GWG).

GWG relative to recommended GWG b	Male–male Twins	Male–female Twins	Female–female Twins	P‐value
Below OR within target range, n (%)	478 (84.45%)	428 (76.98%)	456 (82.01%)	M‐M vs M‐F: 0.026 a M‐F vs F‐F: 0.142 M‐M vs F‐F: 0.439
Exceeded target range, n (%)	88 (15.5%)	128 (23.02%)	100 (17.99%)

^a P ≤ 0.05 considered significant.

^b Based on gravidae's preconception BMI and IOM recommendations ²⁴ for normal (BMI 18.5–24.9 kg/m²), overweight (BMI 25–29.9 kg/m²) and gravidae with obesity (BMI ≥30 kg/m²). Underweight (BMI <18.5 kg/m²) gravidae were excluded given no IOM recommendations for appropriate GWG at this preconception BMI.

As highlighted above, however, twin gestations affected by GDM are a unique population and it is important to ascertain whether the same risk factors for progression to GDMA2 in singletons affect gravidae with MG. Variations between risk factors for GDM diagnoses in twins vs singletons have been noted. GDM MG gravidae were less likely to have a family history of diabetes, unlike their singleton counterparts. ²⁵ Compared with singletons, GDM gravidae with twin gestations were less likely to require pharmacological treatment (ie metformin or insulin). ²⁶ When a diagnosis of GDMA2 was made, gravidae with twins required smaller daily amounts of insulin therapy compared with similarly matched singleton gravidae. ²⁵ Thus, twin gestation pregnancies complicated by GDM are a unique population in which clinical outcomes, risk factors and severity of GDM do not necessarily mirror those of singleton gestations. Currently, there is little evidence to guide the management of GDM in twin gestations. Considering fetal sex pairing once a GDM diagnosis is made in a pregnancy complicated by twins may allow practitioners to identify high‐risk groups within this patient population to provide focused treatment plans and improved resource utilization.

Future studies are needed to identify mechanisms by which maternal glucose is differentially affected by the fetal sex pairing in twins, especially in those gravidae diagnosed with GDM. One hypothesis is that a larger placental mass supports twin gestations, and thus increased amounts of placentally derived factors could drive the development of GDM. ⁴ The placenta provides communication between the mother and the fetus. It modulates the flow of nutrients to the unborn offspring and releases placental hormones that might influence maternal metabolism. Trophoblast respiration in lean women uses glucose, fatty acids and glutamine in roughly equal proportions of energy production. ²⁷ However, by changing the maternal metabolic milieu with comorbidities such as obesity and GDMA2, trophoblast utilization of these substrates occurs in a sexually dimorphic manner. ²⁷ Specific to glucose transport, GLUT receptors actively transport glucose. In human placenta, microvillous membrane GLUT1 expression has been shown to correlate negatively with placental glucose uptake, which correlates positively with placental glucose consumption. ²⁷ Wang et al. demonstrated that glucose dependency is negatively correlated with GLUT1 expression only in the male trophoblast. As such, male fetuses are thought to adopt a growth strategy, whereas female fetuses adopt a preservation strategy.

Murine studies also support the idea of sexually dimorphic placental metabolism. Male mice heterozygous for the GLUT4 gene (an insulin‐sensitive glucose transporter) (G4^+/−) born to wild‐type mothers fed a high‐fat diet were noted to have different fetal hepatic histone modifications compared with maternal high‐fat diet‐fed wild‐type offspring even though GLUT4 is not prominently expressed in the fetal liver. ²⁸ However, GLUT4 is expressed in mouse placenta, suggesting that even though both offspring groups were born from wild‐type mothers fed a high‐fat diet, the male G4^+/− offspring had a different intrauterine milieu from the wild‐type offspring. ²⁸ The fetal hepatic histone modifications observed may be attributed to the haploinsufficiency of the GLUT4 transporter in the placenta in combination with altered maternal nutritional uptake. ²⁸ In fact, the male G4^+/− offspring from mothers fed a high‐fat diet were noted to have a 12‐fold increase in the hepatic gene expression of Slc2a2, the glucose transporter GLUT2, which Vuguin et al. argue may serve as a compensatory adaptation to maintain similar glucose levels to the G4^+/− offspring of female mice fed a control diet. ²⁹ These findings from human trophoblast and murine studies provide a reasonable framework in which fetal sex, intrauterine environment, maternal diet, GLUT receptor expression and altered placental metabolism could affect clinical obstetrical outcomes that warrant further investigation.

Additionally, levels of diabetogenic hormones of human placental lactogen, which is secreted by the placenta into maternal circulation and serves as a physiologic antagonist to insulin, and estrogen have been reported to be higher in pregnancies with female neonates than with male neonates. ³⁰ , ³¹ A previous study found gravidae with GDM were more likely to be associated with dichorionic twins, which have two separate placentas, compared with gravidae with monochorionic twin gestations. ³² In our study, our same‐sex fetal pairs (ie M‐M and F‐F) had relatively equal proportions of MCMA, MCDA and DCDA pairings with no differences in rates of GDM. We identified no difference in rates of GDM in the same‐sex fetal pairs compared with the mixed‐sex DCDA pairs. Further research is needed to assess the levels of these hormones in twin gestations, identify how they may vary with fetal sex pairing, evaluate how they may fluctuate depending upon chronicity, and determine how they could potentially play a role in glycemic control and metabolic dysfunction.

There currently is a lack of sufficiently large, well‐designed studies on hormone levels in twins. ³³ A limited study of 27 sets of twins, including 11 sets of twins of mixed sex, found that both male and female twins (from single‐sex and mixed‐sex pregnancies) had lower cord estrogen concentrations than singletons, but levels did not vary with fetal sex. ³⁴ However, data on the presence or absence of a GDM diagnosis were unavailable. Directed studies of hormone levels in specific subpopulations, such as gravidae with both MG and GDM, are necessary to not only better understand metabolic complications of pregnancy but also help better characterize and understand the development of disease processes potentially related to prenatal estrogen exposure, such as breast cancer. We suggest that more studies to define further the pathophysiology driving fetal sex‐dependent clinical outcomes in twin gestations are necessary to improve and inform maternal‐fetal care.

This study's strengths include using a large, population‐based database that relies on manually extracted data rather than diagnosis codes to eliminate the risk of recall bias. Data collection was performed by trained, multilingual research staff and regular database audits to ensure the validity of the data obtained from interviews. With these methods, we could distinguish between gestational and T2DM, ensuring we minimized any diagnostic misclassifications. Although our study is robust, there are limitations to note. Our study is retrospective and as such, all limitations to such a research design are present. As an observational study, it is important to note that our findings cannot address causality and can only infer associations between outcomes. The mechanisms responsible for any effects of sex difference in the fetus on maternal glucose and insulin metabolism in pregnant persons with twins are yet to be determined and warrant further investigation.

5. CONCLUSION

To our knowledge, this study is unique in that it specifically explores the association between fetal sex pairing in gravidae with twin gestations. We did not observe any relationship between the development of GDM and the fetal sex pairing with equal proportions of gravidae with M‐M, M‐F and F‐F twins developing GDM. However, in gravidae with MG that do develop GDM, a higher proportion of deliveries of F‐F twins was associated with GDMA2 when compared with M‐M or M‐F twins. Further research on the physiology driving this association between two female fetuses and medication‐controlled GDM is warranted.

AUTHOR CONTRIBUTIONS

The authors confirm their contribution to the paper as follows: AMS, KMA: study conception and design. AMS: data collection. AMS, HSH: analysis and interpretation of results. AMS, KMA: writing – draft preparation. All authors reviewed the results and approved the final version of the article.

FUNDING INFORMATION

Components of this work were supported by the National Institute of Health, including NIDDK (1R01 DK128187‐01A1, 6R01DK089201 both to KA) and NICHD Women's Reproductive Health Research (WRHR) (K12 HD103087) of which Dr. Sassin is a scholar.

CONFLICT OF INTEREST STATEMENT

The authors have stated explicitly that there are no conflicts of interest in connection with this article.

Sassin AM, Sangi‐Haghpeykar H, Aagaard KM. Fetal sex and the development of gestational diabetes mellitus in gravidae with multiple gestation pregnancies. Acta Obstet Gynecol Scand. 2023;102:1703‐1710. doi: 10.1111/aogs.14625