Development and Multicenter Validation of a Novel Model for Selective Screening of Gestational Diabetes Mellitus: TheVietnam Gestational Diabetes Mellitus Study

Development and Multicenter Validation of a Novel Model for Selective Screening of Gestational Diabetes Mellitus: TheVietnam Gestational Diabetes Mellitus Study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Development and Multicenter Validation of a Novel Model for Selective Screening of Gestational Diabetes Mellitus: TheVietnam Gestational Diabetes Mellitus Study Nga K. Tran, Thanh N. Cao, Linh V. Pham, Tam D. Lam, Trinh A. T. Vo, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7009949/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 07 Nov, 2025 Read the published version in BMC Pregnancy and Childbirth → Version 1 posted 10 You are reading this latest preprint version Abstract Aims Gestational diabetes mellitus is commonly observed in pregnant women and is associated with an increased risk of adverse outcomes for both mother and child, not only during pregnancy but also in the long term thereafter. The present study aimed to develop a predictive nomogram for gestational diabetes mellitus in pregnant women. Materials and methods This multicenter prospective cohort study enrolled 1,398 pregnant women from five major obstetric hospitals in Vietnam’s Mekong Delta. GDM was diagnosed based on the 2017 American Diabetes Association criteria. Using Bayesian Model Averaging, the optimal prediction model was identified in the primary cohort (n = 978) and used to construct a nomogram for individualized risk estimation. Model performance was validated in an independent cohort (n = 420), with assessment of discrimination (AUC), calibration (Brier score), and clinical utility (decision curve analysis). Results The prevalence of GDM was 18.0% (95% CI: 16.0–20.1). The final model included maternal age (OR per year: 1.09; 95% CI: 1.06–1.13), history of macrosomia (OR: 6.04; 95% CI: 2.76–13.19), body mass index (OR per kg/m²: 1.62; 95% CI: 1.25–2.10), and weight gain during pregnancy (OR per kg: 1.12; 95% CI: 1.06–1.18). The model demonstrated good discriminative ability in the primary cohort (AUC = 0.74, Brier score = 0.123), and acceptable performance in the validation cohort (AUC = 0.70; 95% CI: 0.63–0.77). The nomogram showed good calibration and yielded higher net benefit across a wide range of risk thresholds (0.1–0.4) in decision curve analysis, indicating strong clinical utility. Conclusions A nomogram incorporating four routinely assessed clinical parameters offers good predictive accuracy for gestational diabetes mellitus. This model may facilitate early identification and targeted intervention for high-risk pregnant women in both resource-limited and clinical settings. Gestational diabetes mellitus risk factors predictive model nomogram Figures Figure 1 Figure 2 Figure 3 Figure 4 I/ Introduction Gestational diabetes mellitus (GDM) is defined as diabetes diagnosed in the second or third trimester of pregnancy in women who did not have overt diabetes prior to gestation and who do not meet criteria for other types of diabetes, such as type 1 diabetes [ 1 ]. Globally, the age-standardized prevalence of GDM is estimated at 14.0%; however, this rate may vary significantly depending on differences in risk factors, screening practices, and diagnostic criteria. Notably, the prevalence of GDM is on the rise worldwide, mirroring the increasing trends in obesity and type 2 diabetes [ 2 , 3 ]. In Vietnam, several studies have reported GDM prevalence rates comparable to or even exceeding the global average, ranging from 6.4–27.1% [ 4 , 5 ]. Pregnant women diagnosed with GDM are at a significantly higher risk of adverse outcomes for both mother and child compared to those without the condition. Adjusted analyses have shown that women with GDM are more likely to undergo cesarean delivery and experience preterm birth, macrosomia, low Apgar scores, neonatal respiratory distress syndrome, neonatal jaundice, and admission to neonatal intensive care units—all at statistically significant levels. Furthermore, GDM, especially when insulin treatment is required, has been closely linked to gestational hypertension, increased need for postpartum blood transfusion, and neonatal respiratory support [ 3 , 6 , 7 ]. Beyond the acute complications during pregnancy and the perinatal period, GDM is also considered a key contributor to the future development of type 2 diabetes in both mothers and their offspring [ 8 ]. These factors collectively contribute to the substantial healthcare burden posed by GDM, leading to rising costs for treatment and maternal-child health services [ 9 , 10 ]. In this context, nomogram models have gained attention as a simple, rapid, and equipment-free tool suitable for early screening. In Vietnam, nomograms have been developed for conditions such as chronic kidney disease and osteoporotic fractures [ 11 – 13 ]. However, in obstetrics, especially for predicting GDM, such models remain limited despite the need for early diagnosis and management. This study was therefore conducted to address this gap. Therefore, early screening and diagnosis, followed by appropriate management strategies, are critically important for improving maternal and neonatal outcomes. II/ Methods 1. Study settings and participants This study was a prospective, multi-center cohort study conducted from 2017 to 2022 in Can Tho Gynaecology and Obstetrics Hospital, An Giang Obstetric and Pediatric Hospital, Soc Trang Obstetric and Pediatric Hospital, and Ca Mau Obstetric and Pediatric Hospital. These are major specialized hospitals at the provincial level, located in the Mekong Delta region of Vietnam. To develop a prediction model, we applied Peduzzi’s method to estimate the sample size based on the number of events per predictor variable. According to this method, a minimum ratio of 10 events per predictor is recommended [ 14 ]. The number of predictors included in the model depends on resources, applicability, predictive ability, and the availability of the study. Therefore, we estimated that the GDM prediction model would include 6 predictors, requiring a minimum of 60 GDM events. Based on the study of Nguyen C. L. in Vietnam, the prevalence of GDM is 6.4% according to the American Diabetes Association [ 4 ], corresponding to a minimum of approximately 938 subjects to yield 60 GDM events. Thus, the sample size of 978 individuals in the primary cohort is sufficient to develop the GDM prediction model. To estimate the sample size for the validation cohort, we applied the formula proposed by Obuchowski (2004), in which p represents the prevalence of gestational diabetes mellitus, estimated at 6.4% based on the aforementioned study. The significance level α was also set at 0.05, and the type II error rate β was set at 0.20. The null hypothesis AUC (AUC₀) was set at 0.5, while the desired AUC was set at 0.7. Based on these parameters, we estimated that at least 370 participants would be required to yield a minimum of 24 GDM events. Therefore, the sample size of 420 participants selected for the validation cohort was deemed sufficient to perform external validation. We planned to select samples from two hospitals located in different regions to develop the predicted model, and two other hospitals from two distinct regions would be used for the validation cohort. A multistage sampling method was applied. Stage 1 involves cluster sampling by selecting provinces in the Mekong Delta region of Vietnam with specialized obstetric hospitals. These include: Can Tho City, An Giang, Tra Vinh, Soc Trang, Tien Giang, Ca Mau, and Hau Giang. Among these, Can Tho (a central urban area) was selected as a fixed cluster, while three other provinces were randomly selected from the remaining ones. We identified An Giang, Soc Trang, and Ca Mau through random selection. Specifically, Can Tho and An Giang were assigned to the primary cohort, while Ca Mau and Soc Trang were assigned to the validation cohort. Stage 2 involved purposive sampling within each cluster. The largest specialized obstetric hospital in each selected province was chosen. Stage 3 was conducted by sampling within each hospital. Approximately half of the total sample size was allocated to the primary cohort (Can Tho Gynecology and Obstetrics Hospital and An Giang Obstetrics and Pediatrics Hospital) and the validation cohort (Ca Mau Obstetrics and Pediatrics Hospital and Soc Trang Obstetrics and Pediatrics Hospital). We selected 978 participants for the primary cohort and 420 participants for the validation cohort. Eligible participants included pregnant women with singleton pregnancies who could accurately recall their last menstrual period and/or had a first-trimester ultrasound, agreed to participate in the study, and underwent a 75g oral glucose tolerance test along with blood sampling as per the gestational diabetes mellitus (GDM) screening protocol. Exclusion criteria included individuals who could not complete all three required blood samples, those who conceived through ovulation induction or in vitro fertilization, and women with a pre-existing diagnosis of diabetes mellitus prior to pregnancy. Additionally, women with medical conditions known to affect glucose metabolism—such as hyperthyroidism, hypothyroidism, Cushing’s syndrome, polycystic ovary syndrome, liver disease, renal failure, malignancies, severe internal medicine conditions, cardiovascular diseases, psychiatric disorders—or those using medications that interfere with glucose metabolism—such as corticosteroids, salbutamol, sympatholytic agents, thiazide diuretics, antipsychotic drugs, acetaminophen, phenytoin, or nicotinic acid—were also excluded from the study. The study’s procedure and protocol were approved by the research and ethics committee of University of Medicine and Pharmacy, Hue University, Vietnam (Approval No. 1435/QĐ-ĐHYD, issued on July 30, 2014), and were conducted according to the ethical principles of the Declaration of Helsinki, and all participants gave a written informed consent. 2. Data collection A standardized questionnaire was used to collect demographic and clinical data on early pregnancy. The anthropometric parameters included age, body mass index, weight gain during pregnancy, family history of GDM, history of macrosomia, history of stillbirth, history of congenital abnormalities, history of preterm birth, history of gestational diabetes mellitus, history of miscarriage, and number of pregnancies. Then, participants were followed up until the glucose tolerance test was performed at 24–28 weeks of gestation. 3. Diagnostic criteria The diagnosis of gestational diabetes mellitus was made using the one-step strategy recommended by the American Diabetes Association in 2017, based on the criteria established by the International Association of the Diabetes and Pregnancy Study Groups [ 15 ]. All participants underwent a 75 g oral glucose tolerance test between 24 and 28 weeks of gestation, following at least three days of usual diet and a minimum fasting period of 8 hours. Plasma glucose levels were measured at fasting, 1 hour, and 2 hours after glucose ingestion. A diagnosis of GDM was made if any of the following thresholds were met or exceeded: fasting plasma glucose ≥ 5.1 mmol/L, 1-hour glucose ≥ 10.0 mmol/L, or 2-hour glucose ≥ 8.5 mmol/L. 4. Data and statistical analyses The Bayesian Model Averaging (BMA) method was used to search for the optimal model for predicting GDM, as it has been consistently found to be more robust than the stepwise model-building method in the selection of an optimal prediction model [ 16 , 17 ]. In the presence of m variables, the regression analysis was carried out for 2 m competing models in the BMA. The regression coefficients were averaged over all possible models. A uniform prior probability was given to each model, and together with the likelihood of each model, the posterior probability of the best model was determined by using the Bayesian theorem. The advantages of this method are that it eliminates insignificant variables and reflects the uncertainty of model selection [ 18 ]. The included variables in the BMA analysis were age, body mass index, weight gain during pregnancy, family history of GDM, history of GDM, history of macrosomia, history of stillbirth, history of preterm birth, history of miscarriage, history of congenital abnormalities, and number of pregnancies. The receiver operating characteristic curve analysis and its corresponding area under the curve (AUC) were used to assess the discriminative performance of the prognostic models [ 19 ]. The 95%CI of the AUC was estimated using the bootstrap method with 100 interactions of 10-fold cross-validation samples. The calibration of prognostic models was assessed using the Brier score. We also developed a predictive nomogram to facilitate the implementation of the prediction model in clinical practice using the “ rms ” software package [ 20 ]. The analyses were conducted using the R Statistical Environment [ 21 ]. External validation was performed using an independent dataset to evaluate the model's generalizability further. The discriminative ability and the calibration of the model were respectively assessed by calculating the AUC, along with its 95% confidence interval, and by evaluating the calibration plot in the validation cohort. Additionally, Decision Curve Analysis (DCA) was conducted to examine the model's clinical utility. DCA evaluates the net benefit of the prediction model across a range of threshold probabilities, assessing whether using the model in clinical decision-making provides more benefit than harm compared to treating all or no patients. This analysis was performed using the “rmda” package in R. III/ Results A total of 1398 subjects were recruited and followed until the end of the study. The primary and validation cohort groups presented no significant disparity, with an average age of 28.6 ± 5.82 for the primary group versus 28.93 ± 6.18 for the validation group (p = 0.342). The detailed information is presented in Table 1 . Table 1 Baseline characteristics of 1398 participants stratified by primary and validation cohort Characteristics Primary Cohort (N = 978) Validation Cohort (N = 420) Total (N = 1398) p n (%) n (%) n (%) Age 18-<25 269 (27.5) 102 (24.3) 371 (26.5) 0.405 a ≥ 25 532 (54.4) 234 (55.7) 766 (54.8) ≥ 35 177 (18.1) 84 (20) 261 (18.7) Mean ± SD 28.6 ± 5.82 28.93 ± 6.18 28.7 ± 5.93 0.342 c Family history of GDM Yes 100 (10.2) 52 (12.4) 152 (10.9) 0.235 a No 878 (89.8) 368 (87.6) 1246 (89.1) History of macrosomia Yes 32 (3.3) 7 (1.7) 39 (2.8) 0.095 a No 946 (96.7) 413 (98.3) 1359 (97.2) History of stillbirth Yes 26 (2.7) 8 (1.9) 34 (2.4) 0.402 a No 952 (97.3) 412 (98.1) 1364 (97.6) History of congenital abnormalities Yes 1 (0.1) 1 (0.2) 2 (0.1) 0.511 b No 977 (99.9) 419 (99.8) 1396 (99.9) History of preterm birth Yes 31 (3.2) 4 (1) 35 (2.5) 0.055 a No 947 (96.8) 416 (99) 1363 (97.5) History of GDM Yes 4 (0.4) 1 (0.2) 5 (0.4) 1 b No 974 (99.6) 419 (99.8) 1393 (99.6) History of miscarriage Yes 203 (20.8) 101 (24) 304 (21.7) 0.171 a No 775 (79.2) 319 (76) 1094 (78.3) Number of pregnancies 1 440 (45) 190 (45.2) 630 (45.1) 0.468 a 2 411 (42) 185 (44) 596 (42.6) 3 127 (13) 45 (10.7) 172 (12.3) Mean ± SD 1.68 ± 0.69 1.65 ± 0.66 1.67 ± 0.68 0.527 c Body mass index < 25 868 (88.8) 368 (87.6) 1236 (88.4) 0.544 a ≥ 25 110 (11.2) 52 (12.4) 162 (11.6) Mean ± SD 21.26 ± 3.18 21.33 ± 3.22 21.28 ± 3.19 0.701 c Weight gain Mean ± SD 7.09 ± 3.24 7.42 ± 3.27 7.19 ± 3.25 0.086 c GDM prevalence Yes 170 (17.4) 81 (19.3) 251 (18.0) 0.395 a No 808 (82.6) 339 (80.7) 1147 (82.0) Note : Comparison of the differences are given according to the a Pearson Chi-Square, b Fisher's Exact Test, c Independent Samples Test. Statistical significance: p < 0.05. SD = Standard deviation, GDM = gestational diabetes mellitus. The BMA method was utilized to search for the optimal prediction model with the fewest predictors and maximal predictive performance. There were a total of 3 models selected for presentation (Table 2 ). The results showed that the posterior probability gradually decreases from model I to model III (0.763 to 0.045). In addition, the model I exhibits the lowest BIC and the lowest number of predictors. Table 2 Characteristics of three models with the BMA method Predictor Probability Model I Model II Model III Age 100 0.09 0.09 0.09 Family history of GDM 14.8 - 0.5 History of gestational diabetes 4.3 - - History of macrosomia 100 1.8 1.74 1.8 History of stillbirth 0 - - - History of congenital abnormalities 0 - - - Number of pregnancies 0 - - - History of preterm birth 0 - - - History of miscarriage 4.5 - - 0.24 Body mass index 100 0.15 0.14 0.15 Weight gain during pregnancy 100 0.12 0.11 0.12 BIC -5912 -5909 -5906 Post probability 0.763 0.148 0.045 Models I is the most suitable candidate for a prognostic model for GDM. Specifically, in model I, pregnant woman with a history of macrosomia had a 6.04-fold increase in the odds of developing GDM. Additionally, for each one-year increase in maternal age and each one kg/m 2 increase in body mass index, the odds of GDM increased by 9% and 16%, respectively. Similarly, the odds of GDM increased by 12% for each one kg increase in weight gain (Table 3 ). Table 3 Predictors of the risk of GDM: logistic regression analysis of model I Predictor Unit of Comparison OR 95%CI Model I: Predictors are age, history of macrosomia, body mass index and weight gain Age Yes 5.55 3.31–9.33 History of macrosomia Yes 2.06 1.31–3.22 Body mass index Yes 3.20 1.67–6.13 Weight gain C 2.91 1.84–4.61 Posterior probability 0.763 Analysis of the receiver operating characteristic curve showed that our prediction model had a fairly good discriminative performance in distinguishing individuals with GDM from those without GDM (AUC = 0.74; 95%CI: 0.69–0.79 for model I. The calibration plot also showed that model I had a good degree of calibration, especially given the close alignment of the logistic calibration line with the ideal line. The Brier scores of 0.123 indicate reasonably good predictive performance (Fig. 1). Figure 2 represents a nomogram that can be used to predict the risk of GDM for an individual with a specific risk profile. For example, a 30-year-old pregnant woman with a history of macrosomia, a body mass index of 25 kg/m 2 and a 14 kg gain during pregnancy would have an 80% risk of GDM. Usage instruction: Mark an individual’s age on the “Age” axis and draw a vertical line to the “Point” axis to determine how many points toward the probability of GDM the individual receives for her age value. Repetition is required for each additional risk factor. Sum up the risk factor's salient features. Prolong a vertical line from the final sum on the “Total Points” axis to intersect with the “Risk of GDM” axis to determine the individual's likelihood of sustaining GDM. The overall risk of GDM for a 30-year-old pregnant woman with a history of macrosomia, a body mass index of 25 kg/m 2 , and a 14 kg gain during pregnancy is illustrated as follows: yes at “history of macrosomia” has 56 points, the “age” of 30 has 42 points”; the “BMI” of 25 kg/m 2 has 50 points and the ”weight gain” of 14 has 50 points. As a result, the sum of the individual point values yields 198. Extending a vertical line from the 198-point mark on the “Total Point” axis to intersect with the "Risk of GDM" axis, we can determine a probability of 0.8, equivalent to 80%. When applied to the validation cohort, our prediction model had a fairly good discriminative performance (AUC = 0.7, 95%CI: 0.63–0.77). In addition, the plot in Fig. 3 showed that the nomogram had a good degree of calibration, especially given the close alignment of the logistic calibration line with the ideal line. Finally, the decision curve analysis demonstrated that the nomogram model yielded a higher net benefit across a wide range of threshold probabilities (approximately 0.1–0.4) compared to individual predictors such as age, history of macrosomia, body mass index, and weight gain. In this range, the nomogram consistently outperformed both the “treat-all” and “treat-none” strategies, indicating superior clinical utility in decision-making (detailed in Fig. 4 ). IV/ Discussion The prevalence of gestational diabetes mellitus is increasing and is associated with various complications for both mother and child. Therefore, the development of prognostic tools for early screening is essential to optimize resource allocation and improve the effectiveness of maternal healthcare. The present study successfully developed and fully validated a prognostic model for assessing the risk of GDM among Vietnamese pregnant women. Key findings from the optimal model identified four independent predictors of GDM: increased maternal age (OR = 1.09 per additional year; 95%CI: 1.06–1.13), a history of delivering a macrosomic infant (OR = 6.04; 95%CI: 2.76–13.19), higher pre-pregnancy BMI (OR = 1.16 per unit increase; 95%CI: 1.09–1.23), and gestational weight gain (OR = 1.12 per kilogram gained; 95%CI: 1.06–1.18). A nomogram was developed based on the model and demonstrated strong predictive performance with high accuracy in both the derivation cohort and the independent validation cohort, indicating good stability and potential applicability in clinical practice. In the present study, we identified that advanced maternal age, a history of macrosomia, higher pre-pregnancy BMI, and excessive gestational weight gain were associated with gestational diabetes mellitus. Over the years, high-quality studies have elucidated the relationship between several key risk factors and the development of GDM. There is substantial evidence indicating that the risk of GDM increases with maternal age. While some studies suggest that the lowest risk is observed in women under 25 years of age, others propose a higher threshold (≥ 30–35 years) as the point at which the risk becomes more pronounced. A large meta-analysis revealed a linear increase in GDM risk with advancing maternal age, with women aged 30–34 having a 1.7-fold higher risk, those aged 35–39 a 2.7-fold increase, and women aged ≥ 40 experiencing a 3.5-fold higher risk compared to the 25–29 age group. Notably, each additional year of age after 18 increases GDM risk by approximately 7.9%, with Asian women showing a nearly double rate of increase (12.7%) compared to women of European descent (6.5%). Several studies have also reported the highest GDM prevalence among Asian women, despite their lower average age and BMI compared to White women. Findings from studies in Vietnam have demonstrated that women aged 25–34 have approximately double the risk of GDM, while those aged ≥ 35 have a threefold increased risk compared to women under 25 years of age. Another important predictor of GDM identified in our study was a history of delivering a macrosomic infant. From a pathophysiological perspective, this may reflect underlying maternal glucose intolerance in a previous pregnancy (including undiagnosed GDM), or genetic and metabolic factors that predispose to fetal overgrowth. A 2018 meta-analysis conducted in Asia showed that women with a history of macrosomia had approximately a 4.4-fold increased risk of developing GDM in subsequent pregnancies. This association has been consistently confirmed in multiple cohort studies and large-scale meta-analyses, where a previous delivery of a macrosomic infant (> 4,000 g) significantly increased the risk of GDM—by at least twofold—in the next pregnancy. Recent data from Vietnam also confirm that a history of delivering a macrosomic infant is associated with an increased risk of GDM (OR = 2.5; 95%CI: 1.2–5.3) [ 5 ]. Pre-pregnancy overweight and obesity emerged as one of the strongest risk factors for GDM in our study. Excess adiposity contributes to insulin resistance through increased fat accumulation and hormonal dysregulation, making it more difficult for the maternal body to maintain normal glucose levels during the heightened insulin demands of pregnancy [ 22 – 24 ]. Our findings are supported by previous data across diverse populations. A global meta-analysis showed that women who were overweight or obese before pregnancy had a 2.6-fold increased risk of developing GDM (95%CI: 1.56–4.45) compared to women with a normal BMI [ 25 ]. In studies focusing on Asian populations, the risk was even higher, up to 3.27 times (95%CI: 2.81–3.80) for women with a BMI ≥ 25 compared to those with BMI < 25 [ 26 ]. Similarly, prior research in Vietnam found that overweight women before pregnancy had a 1.6-fold increased risk of GDM compared to women with normal weight [ 5 ]. In Western populations, where obesity rates are higher and the impact of BMI is more pronounced, the risk of GDM among overweight and obese women can increase by 2 to 8 times [ 27 , 28 ]. This trend clearly demonstrates that pre-pregnancy obesity places women at a significantly elevated risk of developing glucose intolerance during pregnancy. In addition to BMI, excessive gestational weight gain (GWG) is also a potential risk factor for GDM. A meta-analysis indicated that women who gained weight above the recommended levels before the typical GDM screening period (24–28 weeks) had a 40% higher risk of developing GDM (OR = 1.40; 95%CI: 1.21–1.61), regardless of their baseline BMI [ 29 ]. This suggests that even women with a normal pre-pregnancy weight may face an increased GDM risk if they gain weight too rapidly or excessively during pregnancy. Multiple studies support this association. A recent large cohort study in China found that excessive weight gain during early pregnancy was associated with a 20% increased risk of GDM [ 30 ]. Other studies have shown that weight gain during the first and second trimesters is more strongly associated with GDM risk than weight gain in the third trimester [ 31 – 33 ]. This highlights the importance of the early gestational window, particularly before 28 weeks, when excessive weight gain can significantly increase glucose intolerance risk. Clearly, excessive GWG—primarily due to abnormal fat accumulation—worsens insulin resistance, making glucose regulation more difficult, especially with early visceral fat deposition during pregnancy [ 29 ]. Additionally, unhealthy dietary patterns and physical inactivity, both often associated with excessive GWG, further contribute to the elevated risk of GDM. Finally, the impact of GWG on GDM risk appears to be consistent across ethnic groups, likely due to shared physiological characteristics of pregnancy. In summary, maternal age, a history of macrosomia, elevated pre-pregnancy BMI, and excessive gestational weight gain are consistent and well-established risk factors for GDM. Pregnant women exhibiting these risk factors should be actively screened for GDM to improve early detection and intervention [ 29 ]. This study successfully developed dedicated prognostic models that allow for rapid and convenient prediction of GDM. Among these, we identified and analyzed three optimal models. Notably, Model 1 fulfilled key criteria for both simplicity and reliability. It comprises only four variables—maternal age, history of macrosomia, pre-pregnancy BMI, and gestational weight gain—all of which are routinely available in standard clinical practice. This model demonstrated superior post-test probability (0.763), the lowest Bayesian Information Criterion (BIC = − 5912), and provided acceptable discriminatory power and good predictive accuracy. The model's discrimination performance, reflected by the area under the ROC curve (AUC), was 0.74 (95%CI: 0.69–0.78). We found our model to be comparable to existing GDM prediction models worldwide. In China—a neighboring country with sociocultural characteristics similar to Vietnam—Zhang et al. reported a simple predictive model including maternal age, BMI, HbA1c, and triglycerides, with an internal AUC of 0.728 (95%CI: 0.683–0.772) [ 34 ]. Gao et al. developed a GDM risk score based on six variables at the first antenatal visit (maternal age, BMI, height, family history of diabetes, systolic blood pressure, and ALT), achieving an internal validation AUC of 0.710. When four lifestyle-related variables were added, the AUC increased only slightly to 0.712 [ 35 ]. Another model by Wu et al. aimed at early GDM prediction (before 16 weeks of gestation) included up to 15 first-trimester parameters and yielded an AUC of 0.746 [ 36 ]. Similarly, several other developed models have demonstrated comparable AUC values in training cohorts [ 37 ]. In Europe, Benhalima et al. proposed a combined clinical model incorporating at least seven variables (including first-degree family history of diabetes, pre-pregnancy smoking, prior GDM, Asian ethnicity, maternal age, height, and pre-pregnancy BMI), achieving an internal AUC of 0.72 (95%CI: 0.69–0.76) [ 38 ]. Notably, the number of variables in these models was equal to or greater than ours, yet their discrimination performance was comparable or inferior. To enhance predictive capacity, several advanced models have incorporated early biochemical markers or utilized machine learning algorithms on large datasets. For example, when Benhalima et al. added biochemical variables, the AUC improved to 0.76 (95%CI: 0.72–0.79) [ 39 ]. In China, a model incorporating adiponectin and PAPP-A reached an AUC of 0.867 [ 40 ]. The MIDO GDM project in Mexico developed a simple artificial neural network with nine easily accessible variables, achieving an AUC of 0.85 [ 41 ]. A deep neural network using 73 clinical and laboratory features reached an AUC of 0.80, later simplified to a seven-variable model with an AUC of approximately 0.77 [ 42 ]. More complex still, a multicenter Korean study using LightGBM/XGBoost algorithms and up to 361 variables achieved an AUC of 0.80 [ 43 ]. While these advanced models demonstrate impressive AUC values, their complexity limits feasibility for iniial screening. More importantly, our model showed good calibration accuracy, as reflected by the Brier score. With a Brier score of 0.123, our model demonstrated well-calibrated risk predictions, a valuable strength that reflects the reliability of its estimated probabilities. In contrast, most other models emphasize discriminatory power and present calibration plots but rarely report Brier scores. The inclusion of a clear and favorable Brier score differentiates our model and highlights its potential for real-world clinical application. In addition, a common limitation of many existing models is that they are primarily evaluated using training datasets, with limited validation in independent populations. Our study’s prediction model demonstrated acceptable discriminative ability in the validation cohort, with an AUC of 0.70 (95%CI: 0.63–0.77). Compared to existing literature, our model achieved similar performance to the model by Guo and colleagues (AUC = 0.70; 95%CI: 0.68–0.72) [ 37 ], and slightly lower than those reported by Wei et al. and Zhang et al. (AUC/C-index ranging approximately from 0.735 to 0.763) [ 44 , 45 ]. In a broader context, Benhalima et al. reported AUCs ranging from 0.717 to 0.769 across multiple external validation cohorts; this model is considered one of the most stable in clinical settings [ 46 ]. Similarly, Meertens et al. reported external validation AUCs for several models ranging from 68–75% [ 47 ]. However, unlike many of these models, which rely on more variables or require complex laboratory measurements, our model is parsimonious and based on easily obtainable clinical parameters. Beyond the AUC, we assessed the model's performance using calibration plots and decision curve analysis. The calibration plot showed a close alignment between the predicted probabilities and the observed outcomes, indicating that the model provides well-calibrated and reliable individual risk estimates. The DCA further demonstrated that our nomogram yielded a greater net clinical benefit than each predictor, particularly within the clinically relevant threshold probability range of 0.1 to 0.4. Other referenced models also showed good agreement between predicted and observed values in their calibration assessments [ 37 ]. While these models demonstrated net clinical benefit across a broader range of threshold probabilities, their use of more numerous or complex variables may limit their feasibility for early screening, especially in primary care or low-resource settings. However, this does not necessarily imply that our model is superior, as models were developed in different populations and have not been directly cross-validated. Our model demonstrates acceptable discrimination, good calibration, and favorable net clinical benefit within a practical decision-making range. We employed Bayesian analytical principles to identify key predictors and select the optimal model for GDM prediction to achieve these outcomes. Using Bayesian Model Averaging allowed us to evaluate all possible models and derive inferences and predictions based on posterior model probabilities. This approach overcomes the limitations of traditional stepwise methods [ 16 ]. As a result, our final model includes a small number of easily obtainable variables, which can be collected early in pregnancy. Importantly, these variables are well-established predictors and are commonly used as standard components in many existing GDM risk models. In summary, using robust statistical techniques such as BMA and internal bootstrap validation supports our model's simplicity, efficiency, and potential for broad clinical application in early GDM screening. The findings of this study hold important implications for the early prevention of gestational diabetes mellitus. We successfully developed a predictive nomogram to assess individualized GDM risk among pregnant women. Notably, our algorithm is applicable not only in low- and middle-income countries but also globally. Given its simplicity and ease of use, the nomogram is expected to be readily implemented in routine clinical practice. As previously discussed, the four predictors—maternal age, history of macrosomia, pre-pregnancy BMI, and gestational weight gain—can be easily obtained through patient history and routine examination. As we know, GDM increases the risk of adverse pregnancy outcomes for both mother and child. Therefore, early identification and prevention in high-risk pregnant women have the potential to reduce such complications and improve overall maternal and neonatal health. The model may be helpful in the early identification of women at high risk of developing GDM, supporting targeted monitoring or preventive interventions while reducing unnecessary procedures in low-risk individuals. Interpreting our research findings requires a balanced consideration of both strengths and limitations. The study used rigorous and standardized methodology to collect a relatively adequate sample size from four independent centers. Model development incorporated internal validation via bootstrap resampling and Bayesian Model Averaging, enhancing robustness. Moreover, model performance was evaluated through discrimination, calibration plots, and decision curve analysis, ensuring comprehensive assessment. Importantly, the model was tested on an independent validation cohort, strengthening the credibility of its predictive performance. However, several limitations warrant attention. First, all study sites were located in large cities within the Mekong Delta region in southern Vietnam, which may restrict the generalizability of findings to the broader national population. Although the model showed an acceptable level of discrimination, its performance remains modest. This is partly attributable to the inherently multifactorial nature of GDM, which makes it challenging to attain substantially higher AUC values in predictive models. Furthermore, the absence of external validation in populations with distinct comorbidities or demographic profiles limits the model's applicability across diverse clinical settings. These limitations underscore the need for future research involving larger, more heterogeneous populations, including those from other geographic regions and healthcare levels. V/ Conclusion Our prospective multi-center cohort study showed that the independent risk factors for gestational diabetes mellitus, including maternal age, pre-pregnancy body mass index, weight gain during pregnancy, and history of macrosomia, were utilized to construct the predictive model. This model demonstrated a fairly good predictive performance and good calibration in both primary and validation cohorts, highlighting its potential as a clinical tool in identifying individuals at risk for gestational diabetes mellitus. Abbreviations AUC – Area Under the Curve BIC – Bayesian Information Criterion BMA – Bayesian Model Averaging BMI – Body Mass Index CI – Confidence Interval DCA – Decision Curve Analysis GDM – Gestational Diabetes Mellitus OGTT – Oral Glucose Tolerance Test OR – Odds Ratio ROC – Receiver Operating Characteristic SD – Standard Deviation Declarations Ethics approval All procedures performed in studies involving human participants were in accordance with the ethical standards of the Ethics Committee of University of Medicine and Pharmacy, Hue University, Vietnam (No. 1435/QĐ-ĐHYD). Consent to participate Informed consent was obtained from all individual participants included in the study. Consent to publish Additional informed consent was obtained from all individual participants for whom identifying information is included in this article. Competing interests The authors have declared that no competing interests exist. Funding The authors have declared that there was no funding for this study. Author Contribution Conception and design of the study: Nga K. Tran, Thanh N. Cao, Linh V. Pham; Acquisition of data, analysis, and interpretation of data: Nga K. Tran, Thanh N. Cao, Linh V. Pham, Tam D. Lam, Trinh A. T. Vo, Thu T. Nguyen, Nghia N. Nguyen, Dang H. Chau, Bao T. Nguyen; Final Approval: Nga K. Tran, Thanh N. Cao, Linh V. Pham, Tam D. Lam, Trinh A. T. Vo, Thu T. Nguyen, Nghia N. Nguyen, Dang H. Chau, Bao T. Nguyen. Acknowledgement We are grateful to express our sincere gratitude to the Rectorate Board of University of Medicine and Pharmacy – Hue University, Can Tho Univesity of Medicine and Pharmacy, Can Tho Gynaecology and Obstetrics Hospital, An Giang Obstetric and Pediatric Hospital, Soc Trang Obstetric and Pediatric Hospital, and Ca Mau Obstetric and Pediatric Hospital for creating favorable conditions for this study to be carried out. Data Availability All data are available from the corresponding author by request. References American Diabetes Association Professional Practice Committee: 2. Diagnosis and Classification of Diabetes: Standards of Care in Diabetes—2025. Diabetes Care 2024, 48(Supplement_1):S27-S49. Wang H, Li N, Chivese T, Werfalli M, Sun H, Yuen L, Hoegfeldt CA, Elise Powe C, Immanuel J, Karuranga S et al : IDF Diabetes Atlas: Estimation of Global and Regional Gestational Diabetes Mellitus Prevalence for 2021 by International Association of Diabetes in Pregnancy Study Group's Criteria. Diabetes Res Clin Pract 2022, 183:109050. Sweeting A, Hannah W, Backman H, Catalano P, Feghali M, Herman WH, Hivert M-F, Immanuel J, Meek C, Oppermann ML et al : Epidemiology and management of gestational diabetes. The Lancet 2024, 404(10448):175-192. Nguyen CL, Lee AH, Minh Pham N, Hoang Nguyen PT, Ha AVV, Khac Chu T, Van Duong D, Thi Duong H, Binns CW: Prevalence and pregnancy outcomes of gestational diabetes mellitus by different international diagnostic criteria: a prospective cohort study in Vietnam. J Matern Fetal Neonatal Med 2020, 33(21):3706-3712. Le HM, Nguyen TA, Nguyen DK, Søndergaard Linde D, Bygbjerg IC, Søndergaard J, Duc Nguyen T, Quang Nguyen B, Thi Dang NA, Nguyen XB et al : Prevalence and risk factors of gestational diabetes mellitus among pregnant women in northern Vietnam: a cross-sectional study. Glob Health Action 2025, 18(1):2460339. Ejaz Z, Azhar Khan A, Sebghat Ullah S, Aamir Hayat M, Maqbool MA, Amin Baig A: The Effects of Gestational Diabetes on Fetus: A Surveillance Study. Cureus 2023, 15(2):e35103. Akinyemi OA, Weldeslase TA, Odusanya E, Akueme NT, Omokhodion OV, Fasokun ME, Makanjuola D, Fakorede M, Ogundipe T: Profiles and Outcomes of Women with Gestational Diabetes Mellitus in the United States. Cureus 2023, 15(7):e41360. Sheiner E: Gestational Diabetes Mellitus: Long-Term Consequences for the Mother and Child Grand Challenge: How to Move on Towards Secondary Prevention? Front Clin Diabetes Healthc 2020, 1:546256. Staynova R, Vasileva E, Yanachkova V: Gestational diabetes mellitus: a growing economic concern. Folia Med (Plovdiv) 2022, 64(5):725-732. Dall TM, Yang W, Gillespie K, Mocarski M, Byrne E, Cintina I, Beronja K, Semilla AP, Iacobucci W, Hogan PF: The Economic Burden of Elevated Blood Glucose Levels in 2017: Diagnosed and Undiagnosed Diabetes, Gestational Diabetes Mellitus, and Prediabetes. Diabetes Care 2019, 42(9):1661-1668. Nguyen HT, Nguyen BT, Tran AV, Nguyen TT, Ngo LH, Vo T, Nhung Thai TH, Mai LD, Tran TS, Nguyen TV et al : A predictive nomogram for selective screening of asymptomatic vertebral fractures: The Vietnam Osteoporosis Study. Osteoporos Sarcopenia 2025, 11(1):9-14. Tran Quang B, Pham Tran P, Nguyen Thanh C, Bui Thi N, Do Dinh T, Tran Quang T, Duong Tuan L, Bui Thi Thuy N, Nguyen Anh N: High incidence of type 2 diabetes in a population with normal range body mass index and individual prediction nomogram in Vietnam. Diabet Med 2022, 39(2):e14680. Vo NH, Pham BV, Nguyen NN, Nguyen BT: A Predictive Nomogram for Selective Screening of Chronic Kidney Disease: A Population-Based Study. Can J Kidney Health Dis 2025, 12:20543581241309979. Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR: A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol 1996, 49(12):1373-1379. American Diabetes Association: 2. Classification and Diagnosis of Diabetes. Diabetes Care 2017, 40(Suppl 1):S11-s24. Wang D, Zhang W, Bakhai A: Comparison of Bayesian model averaging and stepwise methods for model selection in logistic regression. Stat Med 2004, 23(22):3451-3467. Genell A, Nemes S, Steineck G, Dickman PW: Model selection in medical research: a simulation study comparing Bayesian model averaging and stepwise regression. BMC Med Res Methodol 2010, 10:108. Ewout W: Clinical Prediction Models: A Practical Approach to Development, Validation, and Updating. Switzerland: Springer Cham; 2009. Hanley JA, McNeil BJ: A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 1983, 148(3):839-843. Harrell F: Regression Modeling Strategies: With Applications to Linear Models, Logistic and Ordinal Regression, and Survival Analysis. Switzerland: Springer Cham; 2015. R Core Team: A language and environment for statistical computing. Vienna, Austria: Foundation for Statistical Computing; 2006. Plows JF, Stanley JL, Baker PN, Reynolds CM, Vickers MH: The Pathophysiology of Gestational Diabetes Mellitus. Int J Mol Sci 2018, 19(11). Catalano PM: Obesity, insulin resistance, and pregnancy outcome. Reproduction 2010, 140(3):365-371. Salameh Mohammad A, Oniya O, Chamseddine Reem S, Konje Justin C, Pan Y: Maternal Obesity, Gestational Diabetes, and Fetal Macrosomia: An Incidental or a Mechanistic Relationship? Maternal-Fetal Medicine 2023, 05(01):27-30. Zhang Y, Xiao CM, Zhang Y, Chen Q, Zhang XQ, Li XF, Shao RY, Gao YM: Factors Associated with Gestational Diabetes Mellitus: A Meta-Analysis. J Diabetes Res 2021, 2021:6692695. Lee KW, Ching SM, Ramachandran V, Yee A, Hoo FK, Chia YC, Wan Sulaiman WA, Suppiah S, Mohamed MH, Veettil SK: Prevalence and risk factors of gestational diabetes mellitus in Asia: a systematic review and meta-analysis. BMC Pregnancy Childbirth 2018, 18(1):494. Chu SY, Callaghan WM, Kim SY, Schmid CH, Lau J, England LJ, Dietz PM: Maternal obesity and risk of gestational diabetes mellitus. Diabetes Care 2007, 30(8):2070-2076. Shin D, Song WO: Prepregnancy body mass index is an independent risk factor for gestational hypertension, gestational diabetes, preterm labor, and small- and large-for-gestational-age infants. J Matern Fetal Neonatal Med 2015, 28(14):1679-1686. Brunner S, Stecher L, Ziebarth S, Nehring I, Rifas-Shiman SL, Sommer C, Hauner H, von Kries R: Excessive gestational weight gain prior to glucose screening and the risk of gestational diabetes: a meta-analysis. Diabetologia 2015, 58(10):2229-2237. Yin A, Tian F, Wu X, Chen Y, Liu K, Tong J, Guan X, Zhang H, Wu L, Niu J: Excessive gestational weight gain in early pregnancy and insufficient gestational weight gain in middle pregnancy increased risk of gestational diabetes mellitus. Chin Med J (Engl) 2022, 135(9):1057-1063. Cho EH, Hur J, Lee KJ: Early Gestational Weight Gain Rate and Adverse Pregnancy Outcomes in Korean Women. PLoS One 2015, 10(10):e0140376. Sommer C, Mørkrid K, Jenum AK, Sletner L, Mosdøl A, Birkeland KI: Weight gain, total fat gain and regional fat gain during pregnancy and the association with gestational diabetes: a population-based cohort study. Int J Obes (Lond) 2014, 38(1):76-81. MacDonald SC, Bodnar LM, Himes KP, Hutcheon JA: Patterns of Gestational Weight Gain in Early Pregnancy and Risk of Gestational Diabetes Mellitus. Epidemiology 2017, 28(3):419-427. Zhang X, Zhao X, Huo L, Yuan N, Sun J, Du J, Nan M, Ji L: Risk prediction model of gestational diabetes mellitus based on nomogram in a Chinese population cohort study. Sci Rep 2020, 10(1):21223. Gao S, Leng J, Liu H, Wang S, Li W, Wang Y, Hu G, Chan JCN, Yu Z, Zhu H et al : Development and validation of an early pregnancy risk score for the prediction of gestational diabetes mellitus in Chinese pregnant women. BMJ Open Diabetes Res Care 2020, 8(1). Wu Y, Ma S, Wang Y, Chen F, Zhu F, Sun W, Shen W, Zhang J, Chen H: A risk prediction model of gestational diabetes mellitus before 16 gestational weeks in Chinese pregnant women. Diabetes Res Clin Pract 2021, 179:109001. Guo F, Yang S, Zhang Y, Yang X, Zhang C, Fan J: Nomogram for prediction of gestational diabetes mellitus in urban, Chinese, pregnant women. BMC Pregnancy Childbirth 2020, 20(1):43. Benhalima K, Van Crombrugge P, Moyson C, Verhaeghe J, Vandeginste S, Verlaenen H, Vercammen C, Maes T, Dufraimont E, De Block C et al : Characteristics and pregnancy outcomes across gestational diabetes mellitus subtypes based on insulin resistance. Diabetologia 2019, 62(11):2118-2128. Benhalima K, Van Crombrugge P, Moyson C, Verhaeghe J, Vandeginste S, Verlaenen H, Vercammen C, Maes T, Dufraimont E, De Block C et al : Estimating the risk of gestational diabetes mellitus based on the 2013 WHO criteria: a prediction model based on clinical and biochemical variables in early pregnancy. Acta Diabetol 2020, 57(6):661-671. Wang X, Sheng Y, Xiao J, Hu Y, Li L, Chen K: Combined detection of serum adiponectin and pregnancy-associated plasma protein A for early prediction of gestational diabetes mellitus. Medicine (Baltimore) 2024, 103(42):e40091. Gallardo-Rincón H, Ríos-Blancas MJ, Ortega-Montiel J, Montoya A, Martinez-Juarez LA, Lomelín-Gascón J, Saucedo-Martínez R, Mújica-Rosales R, Galicia-Hernández V, Morales-Juárez L et al : MIDO GDM: an innovative artificial intelligence-based prediction model for the development of gestational diabetes in Mexican women. Sci Rep 2023, 13(1):6992. Wu YT, Zhang CJ, Mol BW, Kawai A, Li C, Chen L, Wang Y, Sheng JZ, Fan JX, Shi Y et al : Early Prediction of Gestational Diabetes Mellitus in the Chinese Population via Advanced Machine Learning. J Clin Endocrinol Metab 2021, 106(3):e1191-e1205. Kang BS, Lee SU, Hong S, Choi SK, Shin JE, Wie JH, Jo YS, Kim YH, Kil K, Chung YH et al : Prediction of gestational diabetes mellitus in Asian women using machine learning algorithms. Sci Rep 2023, 13(1):13356. Zhang J, Cao Q, Mao C, Xu J, Li Y, Mu Y, Huang G, Chen D, Deng X, Xu T et al : Development and validation of a prediction model for gestational diabetes mellitus risk among women from 8 to 14 weeks of gestation in Western China. BMC Pregnancy Childbirth 2025, 25(1):385. Wei Y, He A, Tang C, Liu H, Li L, Yang X, Wang X, Shen F, Liu J, Li J et al : Risk prediction models of gestational diabetes mellitus before 16 gestational weeks. BMC Pregnancy Childbirth 2022, 22(1):889. Kotzaeridi G, Blätter J, Eppel D, Rosicky I, Mittlböck M, Yerlikaya-Schatten G, Schatten C, Husslein P, Eppel W, Huhn EA et al : Performance of early risk assessment tools to predict the later development of gestational diabetes. Eur J Clin Invest 2021, 51(12):e13630. Meertens LJE, Scheepers HCJ, van Kuijk SMJ, Roeleveld N, Aardenburg R, van Dooren IMA, Langenveld J, Zwaan IM, Spaanderman MEA, van Gelder M et al : External validation and clinical utility of prognostic prediction models for gestational diabetes mellitus: A prospective cohort study. Acta Obstet Gynecol Scand 2020, 99(7):891-900. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 07 Nov, 2025 Read the published version in BMC Pregnancy and Childbirth → Version 1 posted Editorial decision: Revision requested 11 Aug, 2025 Reviews received at journal 05 Aug, 2025 Reviewers agreed at journal 29 Jul, 2025 Reviews received at journal 22 Jul, 2025 Reviewers agreed at journal 22 Jul, 2025 Reviewers invited by journal 09 Jul, 2025 Editor invited by journal 01 Jul, 2025 Editor assigned by journal 30 Jun, 2025 Submission checks completed at journal 30 Jun, 2025 First submitted to journal 30 Jun, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7009949","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":483983871,"identity":"200af31c-f3a3-4997-abf9-aa7d4853adf4","order_by":0,"name":"Nga K. Tran","email":"","orcid":"","institution":"Can Tho University of Medicine and Pharmacy","correspondingAuthor":false,"prefix":"","firstName":"Nga","middleName":"K.","lastName":"Tran","suffix":""},{"id":483983872,"identity":"3a33a832-0ff5-43e7-a3f1-21d76466a786","order_by":1,"name":"Thanh N. Cao","email":"","orcid":"","institution":"University of Medicine and Pharmacy, Hue University","correspondingAuthor":false,"prefix":"","firstName":"Thanh","middleName":"N.","lastName":"Cao","suffix":""},{"id":483983874,"identity":"049a4fd6-3910-46c9-912a-b968aac1845f","order_by":2,"name":"Linh V. Pham","email":"","orcid":"","institution":"Hong Bang International University","correspondingAuthor":false,"prefix":"","firstName":"Linh","middleName":"V.","lastName":"Pham","suffix":""},{"id":483983875,"identity":"66b9d0c0-aa8a-450d-a4f2-5ab52a8514a9","order_by":3,"name":"Tam D. Lam","email":"","orcid":"","institution":"Can Tho University of Medicine and Pharmacy","correspondingAuthor":false,"prefix":"","firstName":"Tam","middleName":"D.","lastName":"Lam","suffix":""},{"id":483983878,"identity":"17fde649-5b24-456f-ae11-b36efce99057","order_by":4,"name":"Trinh A. T. Vo","email":"","orcid":"","institution":"Can Tho University of Medicine and Pharmacy","correspondingAuthor":false,"prefix":"","firstName":"Trinh","middleName":"A. T.","lastName":"Vo","suffix":""},{"id":483983879,"identity":"9a5ab1d7-a40e-47ec-9c10-5fca0598f327","order_by":5,"name":"Thu T. Nguyen","email":"","orcid":"","institution":"Can Tho University of Medicine and Pharmacy","correspondingAuthor":false,"prefix":"","firstName":"Thu","middleName":"T.","lastName":"Nguyen","suffix":""},{"id":483983880,"identity":"fada6611-0d0a-401a-8921-5456ea467ebe","order_by":6,"name":"Nghia N. Nguyen","email":"","orcid":"","institution":"Can Tho University of Medicine and Pharmacy","correspondingAuthor":false,"prefix":"","firstName":"Nghia","middleName":"N.","lastName":"Nguyen","suffix":""},{"id":483983881,"identity":"df556b43-734d-4ca9-9700-fecd7fa36e9f","order_by":7,"name":"Dang H. Chau","email":"","orcid":"","institution":"Hong Bang International University","correspondingAuthor":false,"prefix":"","firstName":"Dang","middleName":"H.","lastName":"Chau","suffix":""},{"id":483983882,"identity":"1a65dc3a-6205-4764-86a9-38c091b1cd50","order_by":8,"name":"Bao T. Nguyen","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA+UlEQVRIiWNgGAWjYDCCAwwMzECKh5+/If3DByCLjZ2wFsZmICUjOePAM8YZIC3MRGqxMTiQ+IyZByRCSAvf7bPHHxfU2PAwHDic9tjm1zZ5PmYGxg8fc3BrkTyXl9g841gaD2NzW7pxbt9twzZmBmbJmdtwazE4w2PYzMN2mIeZ4UyCdG7PbUagFjZmXoJa/h3mYWPI/yBt2XPbnjgtvG2HeXgYEtKkGX7cTiSoRRKoZfbMvjQeCYkDyYa9DbeT25gZm/H6he8Mj8Hngm829vbnGxIf/Phz23Z+e/PBDx/xaEEFjG1gsoFY9SDwhxTFo2AUjIJRMFIAAF77UqimNCgQAAAAAElFTkSuQmCC","orcid":"","institution":"University of Medicine and Pharmacy, Hue University","correspondingAuthor":true,"prefix":"","firstName":"Bao","middleName":"T.","lastName":"Nguyen","suffix":""}],"badges":[],"createdAt":"2025-06-30 11:23:08","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7009949/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7009949/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s12884-025-08249-w","type":"published","date":"2025-11-07T15:58:02+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":86671116,"identity":"e4934e7f-65e3-4805-b289-93242a8a1c82","added_by":"auto","created_at":"2025-07-14 11:35:40","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":825729,"visible":true,"origin":"","legend":"\u003cp\u003eDiscriminative value expressed through the AUC (left) and the calibration of model I (right)\u003c/p\u003e","description":"","filename":"Figure1.tif.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7009949/v1/c481932f8d0241d91dbe2d31.jpg"},{"id":86672885,"identity":"b956e443-bf3d-42ff-a8bb-2f51f710db5a","added_by":"auto","created_at":"2025-07-14 11:43:40","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1052664,"visible":true,"origin":"","legend":"\u003cp\u003eNomogram for predicting the individual risk of gestational diabetes mellitus\u003c/p\u003e","description":"","filename":"Figure2.tif.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7009949/v1/376846d6eb3ceba378174691.jpg"},{"id":86671119,"identity":"ac95d3c7-f98d-4c30-b0c9-342825cbacfb","added_by":"auto","created_at":"2025-07-14 11:35:40","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":457697,"visible":true,"origin":"","legend":"\u003cp\u003eDiscriminative value expressed through the AUC (left) and the calibration of model I (right) inthevalidation cohort\u003c/p\u003e","description":"","filename":"Figure3.tif.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7009949/v1/1870a6a6ed30c5ac7b75a7ca.jpg"},{"id":86671118,"identity":"ea0af9b2-6bdf-4da5-98bf-5259bc445deb","added_by":"auto","created_at":"2025-07-14 11:35:40","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":527422,"visible":true,"origin":"","legend":"\u003cp\u003eThe decision curve analysis for the nomogram in the validation cohort\u003c/p\u003e","description":"","filename":"Figure4.tif.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7009949/v1/23dd2093527d4b192529fbd6.jpg"},{"id":95564295,"identity":"7c012045-a1fb-4cdc-9bcf-37255bb64f82","added_by":"auto","created_at":"2025-11-10 16:09:44","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3728157,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7009949/v1/ea84230d-f122-45ef-a54a-244a2267525e.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"\u003cp\u003eDevelopment and Multicenter Validation of a Novel Model for Selective Screening of Gestational Diabetes Mellitus: TheVietnam Gestational Diabetes Mellitus Study\u003c/p\u003e","fulltext":[{"header":"I/ Introduction","content":"\u003cdiv id=\"Sec2\" class=\"Section2\"\u003e\u003cp\u003eGestational diabetes mellitus (GDM) is defined as diabetes diagnosed in the second or third trimester of pregnancy in women who did not have overt diabetes prior to gestation and who do not meet criteria for other types of diabetes, such as type 1 diabetes [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. Globally, the age-standardized prevalence of GDM is estimated at 14.0%; however, this rate may vary significantly depending on differences in risk factors, screening practices, and diagnostic criteria. Notably, the prevalence of GDM is on the rise worldwide, mirroring the increasing trends in obesity and type 2 diabetes [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. In Vietnam, several studies have reported GDM prevalence rates comparable to or even exceeding the global average, ranging from 6.4–27.1% [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e].\u003c/p\u003e\u003cp\u003ePregnant women diagnosed with GDM are at a significantly higher risk of adverse outcomes for both mother and child compared to those without the condition. Adjusted analyses have shown that women with GDM are more likely to undergo cesarean delivery and experience preterm birth, macrosomia, low Apgar scores, neonatal respiratory distress syndrome, neonatal jaundice, and admission to neonatal intensive care units—all at statistically significant levels. Furthermore, GDM, especially when insulin treatment is required, has been closely linked to gestational hypertension, increased need for postpartum blood transfusion, and neonatal respiratory support [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. Beyond the acute complications during pregnancy and the perinatal period, GDM is also considered a key contributor to the future development of type 2 diabetes in both mothers and their offspring [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. These factors collectively contribute to the substantial healthcare burden posed by GDM, leading to rising costs for treatment and maternal-child health services [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. In this context, nomogram models have gained attention as a simple, rapid, and equipment-free tool suitable for early screening. In Vietnam, nomograms have been developed for conditions such as chronic kidney disease and osteoporotic fractures [\u003cspan additionalcitationids=\"CR12\" citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e–\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. However, in obstetrics, especially for predicting GDM, such models remain limited despite the need for early diagnosis and management. This study was therefore conducted to address this gap. Therefore, early screening and diagnosis, followed by appropriate management strategies, are critically important for improving maternal and neonatal outcomes.\u003c/p\u003e\u003c/div\u003e"},{"header":"II/ Methods","content":"\u003ch2\u003e1. Study settings and participants\u003c/h2\u003e\u003cp\u003eThis study was a prospective, multi-center cohort study conducted from 2017 to 2022 in Can Tho Gynaecology and Obstetrics Hospital, An Giang Obstetric and Pediatric Hospital, Soc Trang Obstetric and Pediatric Hospital, and Ca Mau Obstetric and Pediatric Hospital. These are major specialized hospitals at the provincial level, located in the Mekong Delta region of Vietnam.\u003c/p\u003e\u003cp\u003eTo develop a prediction model, we applied Peduzzi’s method to estimate the sample size based on the number of events per predictor variable. According to this method, a minimum ratio of 10 events per predictor is recommended [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. The number of predictors included in the model depends on resources, applicability, predictive ability, and the availability of the study. Therefore, we estimated that the GDM prediction model would include 6 predictors, requiring a minimum of 60 GDM events. Based on the study of Nguyen C. L. in Vietnam, the prevalence of GDM is 6.4% according to the American Diabetes Association [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e], corresponding to a minimum of approximately 938 subjects to yield 60 GDM events. Thus, the sample size of 978 individuals in the primary cohort is sufficient to develop the GDM prediction model. To estimate the sample size for the validation cohort, we applied the formula proposed by Obuchowski (2004), in which p represents the prevalence of gestational diabetes mellitus, estimated at 6.4% based on the aforementioned study. The significance level α was also set at 0.05, and the type II error rate β was set at 0.20. The null hypothesis AUC (AUC₀) was set at 0.5, while the desired AUC was set at 0.7. Based on these parameters, we estimated that at least 370 participants would be required to yield a minimum of 24 GDM events. Therefore, the sample size of 420 participants selected for the validation cohort was deemed sufficient to perform external validation.\u003c/p\u003e\u003cp\u003eWe planned to select samples from two hospitals located in different regions to develop the predicted model, and two other hospitals from two distinct regions would be used for the validation cohort. A multistage sampling method was applied. Stage 1 involves cluster sampling by selecting provinces in the Mekong Delta region of Vietnam with specialized obstetric hospitals. These include: Can Tho City, An Giang, Tra Vinh, Soc Trang, Tien Giang, Ca Mau, and Hau Giang. Among these, Can Tho (a central urban area) was selected as a fixed cluster, while three other provinces were randomly selected from the remaining ones. We identified An Giang, Soc Trang, and Ca Mau through random selection. Specifically, Can Tho and An Giang were assigned to the primary cohort, while Ca Mau and Soc Trang were assigned to the validation cohort. Stage 2 involved purposive sampling within each cluster. The largest specialized obstetric hospital in each selected province was chosen. Stage 3 was conducted by sampling within each hospital. Approximately half of the total sample size was allocated to the primary cohort (Can Tho Gynecology and Obstetrics Hospital and An Giang Obstetrics and Pediatrics Hospital) and the validation cohort (Ca Mau Obstetrics and Pediatrics Hospital and Soc Trang Obstetrics and Pediatrics Hospital). We selected 978 participants for the primary cohort and 420 participants for the validation cohort.\u003c/p\u003e\u003cp\u003eEligible participants included pregnant women with singleton pregnancies who could accurately recall their last menstrual period and/or had a first-trimester ultrasound, agreed to participate in the study, and underwent a 75g oral glucose tolerance test along with blood sampling as per the gestational diabetes mellitus (GDM) screening protocol. Exclusion criteria included individuals who could not complete all three required blood samples, those who conceived through ovulation induction or in vitro fertilization, and women with a pre-existing diagnosis of diabetes mellitus prior to pregnancy. Additionally, women with medical conditions known to affect glucose metabolism—such as hyperthyroidism, hypothyroidism, Cushing’s syndrome, polycystic ovary syndrome, liver disease, renal failure, malignancies, severe internal medicine conditions, cardiovascular diseases, psychiatric disorders—or those using medications that interfere with glucose metabolism—such as corticosteroids, salbutamol, sympatholytic agents, thiazide diuretics, antipsychotic drugs, acetaminophen, phenytoin, or nicotinic acid—were also excluded from the study.\u003c/p\u003e\u003cp\u003e The study’s procedure and protocol were approved by the research and ethics committee of University of Medicine and Pharmacy, Hue University, Vietnam (Approval No. 1435/QĐ-ĐHYD, issued on July 30, 2014), and were conducted according to the ethical principles of the Declaration of Helsinki, and all participants gave a written informed consent.\u003c/p\u003e\u003ch3\u003e2. Data collection\u003c/h3\u003e\u003cp\u003eA standardized questionnaire was used to collect demographic and clinical data on early pregnancy. The anthropometric parameters included age, body mass index, weight gain during pregnancy, family history of GDM, history of macrosomia, history of stillbirth, history of congenital abnormalities, history of preterm birth, history of gestational diabetes mellitus, history of miscarriage, and number of pregnancies. Then, participants were followed up until the glucose tolerance test was performed at 24–28 weeks of gestation.\u003c/p\u003e\u003ch3\u003e3. Diagnostic criteria\u003c/h3\u003e\u003cp\u003eThe diagnosis of gestational diabetes mellitus was made using the one-step strategy recommended by the American Diabetes Association in 2017, based on the criteria established by the International Association of the Diabetes and Pregnancy Study Groups [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. All participants underwent a 75 g oral glucose tolerance test between 24 and 28 weeks of gestation, following at least three days of usual diet and a minimum fasting period of 8 hours. Plasma glucose levels were measured at fasting, 1 hour, and 2 hours after glucose ingestion. A diagnosis of GDM was made if any of the following thresholds were met or exceeded: fasting plasma glucose ≥ 5.1 mmol/L, 1-hour glucose ≥ 10.0 mmol/L, or 2-hour glucose ≥ 8.5 mmol/L.\u003c/p\u003e\u003ch3\u003e4. Data and statistical analyses\u003c/h3\u003e\u003cp\u003eThe Bayesian Model Averaging (BMA) method was used to search for the optimal model for predicting GDM, as it has been consistently found to be more robust than the stepwise model-building method in the selection of an optimal prediction model [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. In the presence of \u003cem\u003em\u003c/em\u003e variables, the regression analysis was carried out for 2\u003csup\u003em\u003c/sup\u003e competing models in the BMA. The regression coefficients were averaged over all possible models. A uniform prior probability was given to each model, and together with the likelihood of each model, the posterior probability of the best model was determined by using the Bayesian theorem. The advantages of this method are that it eliminates insignificant variables and reflects the uncertainty of model selection [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. The included variables in the BMA analysis were age, body mass index, weight gain during pregnancy, family history of GDM, history of GDM, history of macrosomia, history of stillbirth, history of preterm birth, history of miscarriage, history of congenital abnormalities, and number of pregnancies.\u003c/p\u003e\u003cp\u003eThe receiver operating characteristic curve analysis and its corresponding area under the curve (AUC) were used to assess the discriminative performance of the prognostic models [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. The 95%CI of the AUC was estimated using the bootstrap method with 100 interactions of 10-fold cross-validation samples. The calibration of prognostic models was assessed using the Brier score. We also developed a predictive nomogram to facilitate the implementation of the prediction model in clinical practice using the “\u003cem\u003erms\u003c/em\u003e” software package [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. The analyses were conducted using the R Statistical Environment [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. External validation was performed using an independent dataset to evaluate the model's generalizability further. The discriminative ability and the calibration of the model were respectively assessed by calculating the AUC, along with its 95% confidence interval, and by evaluating the calibration plot in the validation cohort. Additionally, Decision Curve Analysis (DCA) was conducted to examine the model's clinical utility. DCA evaluates the net benefit of the prediction model across a range of threshold probabilities, assessing whether using the model in clinical decision-making provides more benefit than harm compared to treating all or no patients. This analysis was performed using the “rmda” package in R.\u003c/p\u003e"},{"header":"III/ Results","content":"\u003cp\u003eA total of 1398 subjects were recruited and followed until the end of the study. The primary and validation cohort groups presented no significant disparity, with an average age of 28.6 ± 5.82 for the primary group versus 28.93 ± 6.18 for the validation group (p = 0.342). The detailed information is presented in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cdiv class=\"gridtable\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eBaseline characteristics of 1398 participants stratified by primary and validation cohort\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"6\"\u003e\u003c/colgroup\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colspan=\"2\" morerows=\"1\" nameend=\"c2\" namest=\"c1\" rowspan=\"2\"\u003e\u003cp\u003eCharacteristics\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003ePrimary Cohort\u003c/p\u003e\u003cp\u003e(N = 978)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eValidation Cohort\u003c/p\u003e\u003cp\u003e(N = 420)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTotal\u003c/p\u003e\u003cp\u003e(N = 1398)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003ep\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003en (%)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003en (%)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003en (%)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e\u003cp\u003eAge\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e18-\u0026lt;25\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e269 (27.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e102 (24.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e371 (26.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\" morerows=\"2\" rowspan=\"3\"\u003e\u003cp\u003e0.405\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e≥ 25\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e532 (54.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e234 (55.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e766 (54.8)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e≥ 35\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e177 (18.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e84 (20)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e261 (18.7)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMean ± SD\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e28.6 ± 5.82\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e28.93 ± 6.18\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e28.7 ± 5.93\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.342\u003csup\u003ec\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eFamily history of GDM\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e100 (10.2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e52 (12.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e152 (10.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e0.235\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eNo\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e878 (89.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e368 (87.6)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e1246 (89.1)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eHistory of macrosomia\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e32 (3.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e7 (1.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e39 (2.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e0.095\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eNo\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e946 (96.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e413 (98.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e1359 (97.2)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eHistory of stillbirth\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e26 (2.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e8 (1.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e34 (2.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e0.402\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eNo\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e952 (97.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e412 (98.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e1364 (97.6)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eHistory of congenital\u003c/p\u003e\u003cp\u003eabnormalities\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e1 (0.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1 (0.2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e2 (0.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e0.511\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eNo\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e977 (99.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e419 (99.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e1396 (99.9)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eHistory of preterm birth\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e31 (3.2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e4 (1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e35 (2.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e0.055\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eNo\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e947 (96.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e416 (99)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e1363 (97.5)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eHistory of GDM\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e4 (0.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1 (0.2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e5 (0.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e1\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eNo\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e974 (99.6)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e419 (99.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e1393 (99.6)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eHistory of miscarriage\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e203 (20.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e101 (24)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e304 (21.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e0.171\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eNo\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e775 (79.2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e319 (76)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e1094 (78.3)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e\u003cp\u003eNumber of pregnancies\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e440 (45)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e190 (45.2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e630 (45.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\" morerows=\"2\" rowspan=\"3\"\u003e\u003cp\u003e0.468\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e411 (42)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e185 (44)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e596 (42.6)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e127 (13)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e45 (10.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e172 (12.3)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMean ± SD\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e1.68 ± 0.69\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1.65 ± 0.66\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e1.67 ± 0.68\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.527\u003csup\u003ec\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e\u003cp\u003eBody mass index\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u0026lt; 25\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e868 (88.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e368 (87.6)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e1236 (88.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e0.544\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e≥ 25\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e110 (11.2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e52 (12.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e162 (11.6)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMean ± SD\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e21.26 ± 3.18\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e21.33 ± 3.22\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e21.28 ± 3.19\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.701\u003csup\u003ec\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWeight gain\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMean ± SD\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e7.09 ± 3.24\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e7.42 ± 3.27\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e7.19 ± 3.25\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.086\u003csup\u003ec\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eGDM prevalence\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e170 (17.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e81 (19.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e251 (18.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e0.395\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eNo\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e808 (82.6)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e339 (80.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e1147 (82.0)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"6\" nameend=\"c6\" namest=\"c1\"\u003e\u003cp\u003e\u003cem\u003eNote\u003c/em\u003e: Comparison of the differences are given according to the \u003csup\u003ea\u003c/sup\u003ePearson Chi-Square, \u003csup\u003eb\u003c/sup\u003eFisher's Exact Test, \u003csup\u003ec\u003c/sup\u003eIndependent Samples Test. Statistical significance: p \u0026lt; 0.05.\u003c/p\u003e\u003cp\u003eSD = Standard deviation, GDM = gestational diabetes mellitus.\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/table\u003e\u003c/div\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThe BMA method was utilized to search for the optimal prediction model with the fewest predictors and maximal predictive performance. There were a total of 3 models selected for presentation (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). The results showed that the posterior probability gradually decreases from model I to model III (0.763 to 0.045). In addition, the model I exhibits the lowest BIC and the lowest number of predictors.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cdiv class=\"gridtable\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eCharacteristics of three models with the BMA method\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"5\"\u003e\u003c/colgroup\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePredictor\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eProbability\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eModel I\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eModel II\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eModel III\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAge\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e100\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.09\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.09\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.09\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFamily history of GDM\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e14.8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eHistory of gestational diabetes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e4.3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eHistory of macrosomia\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e100\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e1.8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1.74\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e1.8\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eHistory of stillbirth\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eHistory of congenital abnormalities\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eNumber of pregnancies\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eHistory of preterm birth\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eHistory of miscarriage\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e4.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.24\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBody mass index\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e100\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.15\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.14\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.15\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWeight gain during pregnancy\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e100\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.12\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.11\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.12\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eBIC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-5912\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-5909\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-5906\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003ePost probability\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.763\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.148\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.045\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/table\u003e\u003c/div\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eModels I is the most suitable candidate for a prognostic model for GDM. Specifically, in model I, pregnant woman with a history of macrosomia had a 6.04-fold increase in the odds of developing GDM. Additionally, for each one-year increase in maternal age and each one kg/m\u003csup\u003e2\u003c/sup\u003e increase in body mass index, the odds of GDM increased by 9% and 16%, respectively. Similarly, the odds of GDM increased by 12% for each one kg increase in weight gain (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cdiv class=\"gridtable\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003ePredictors of the risk of GDM: logistic regression analysis of model I\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"4\"\u003e\u003c/colgroup\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePredictor\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eUnit of Comparison\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eOR\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003e95%CI\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"4\" nameend=\"c4\" namest=\"c1\"\u003e\u003cp\u003eModel I: Predictors are age, history of macrosomia, body mass index and weight gain\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAge\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e5.55\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e3.31–9.33\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eHistory of macrosomia\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e2.06\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1.31–3.22\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBody mass index\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYes\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e3.20\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1.67–6.13\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWeight gain\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e2.91\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1.84–4.61\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePosterior probability\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"3\" nameend=\"c4\" namest=\"c2\"\u003e\u003cp\u003e0.763\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/table\u003e\u003c/div\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eAnalysis of the receiver operating characteristic curve showed that our prediction model had a fairly good discriminative performance in distinguishing individuals with GDM from those without GDM (AUC = 0.74; 95%CI: 0.69–0.79 for model I. The calibration plot also showed that model I had a good degree of calibration, especially given the close alignment of the logistic calibration line with the ideal line. The Brier scores of 0.123 indicate reasonably good predictive performance (Fig.\u0026nbsp;1).\u003c/p\u003e\u003cp\u003e\u003cb\u003eFigure 2\u003c/b\u003e represents a nomogram that can be used to predict the risk of GDM for an individual with a specific risk profile. For example, a 30-year-old pregnant woman with a history of macrosomia, a body mass index of 25 kg/m\u003csup\u003e2\u003c/sup\u003e and a 14 kg gain during pregnancy would have an 80% risk of GDM.\u003c/p\u003e\u003cp\u003eUsage instruction: Mark an individual’s age on the “Age” axis and draw a vertical line to the “Point” axis to determine how many points toward the probability of GDM the individual receives for her age value. Repetition is required for each additional risk factor. Sum up the risk factor's salient features. Prolong a vertical line from the final sum on the “Total Points” axis to intersect with the “Risk of GDM” axis to determine the individual's likelihood of sustaining GDM. The overall risk of GDM for a 30-year-old pregnant woman with a history of macrosomia, a body mass index of 25 kg/m\u003csup\u003e2\u003c/sup\u003e, and a 14 kg gain during pregnancy is illustrated as follows: yes at “history of macrosomia” has 56 points, the “age” of 30 has 42 points”; the “BMI” of 25 kg/m\u003csup\u003e2\u003c/sup\u003e has 50 points and the ”weight gain” of 14 has 50 points. As a result, the sum of the individual point values yields 198. Extending a vertical line from the 198-point mark on the “Total Point” axis to intersect with the \"Risk of GDM\" axis, we can determine a probability of 0.8, equivalent to 80%.\u003c/p\u003e\u003cp\u003eWhen applied to the validation cohort, our prediction model had a fairly good discriminative performance (AUC = 0.7, 95%CI: 0.63–0.77). In addition, the plot in \u003cb\u003eFig.\u0026nbsp;3\u003c/b\u003e showed that the nomogram had a good degree of calibration, especially given the close alignment of the logistic calibration line with the ideal line.\u003c/p\u003e\u003cp\u003eFinally, the decision curve analysis demonstrated that the nomogram model yielded a higher net benefit across a wide range of threshold probabilities (approximately 0.1–0.4) compared to individual predictors such as age, history of macrosomia, body mass index, and weight gain. In this range, the nomogram consistently outperformed both the “treat-all” and “treat-none” strategies, indicating superior clinical utility in decision-making (detailed in \u003cb\u003eFig.\u0026nbsp;4\u003c/b\u003e).\u003c/p\u003e"},{"header":"IV/ Discussion","content":"\u003cp\u003eThe prevalence of gestational diabetes mellitus is increasing and is associated with various complications for both mother and child. Therefore, the development of prognostic tools for early screening is essential to optimize resource allocation and improve the effectiveness of maternal healthcare. The present study successfully developed and fully validated a prognostic model for assessing the risk of GDM among Vietnamese pregnant women. Key findings from the optimal model identified four independent predictors of GDM: increased maternal age (OR = 1.09 per additional year; 95%CI: 1.06–1.13), a history of delivering a macrosomic infant (OR = 6.04; 95%CI: 2.76–13.19), higher pre-pregnancy BMI (OR = 1.16 per unit increase; 95%CI: 1.09–1.23), and gestational weight gain (OR = 1.12 per kilogram gained; 95%CI: 1.06–1.18). A nomogram was developed based on the model and demonstrated strong predictive performance with high accuracy in both the derivation cohort and the independent validation cohort, indicating good stability and potential applicability in clinical practice.\u003c/p\u003e\u003cp\u003eIn the present study, we identified that advanced maternal age, a history of macrosomia, higher pre-pregnancy BMI, and excessive gestational weight gain were associated with gestational diabetes mellitus. Over the years, high-quality studies have elucidated the relationship between several key risk factors and the development of GDM. There is substantial evidence indicating that the risk of GDM increases with maternal age. While some studies suggest that the lowest risk is observed in women under 25 years of age, others propose a higher threshold (≥ 30–35 years) as the point at which the risk becomes more pronounced. A large meta-analysis revealed a linear increase in GDM risk with advancing maternal age, with women aged 30–34 having a 1.7-fold higher risk, those aged 35–39 a 2.7-fold increase, and women aged ≥ 40 experiencing a 3.5-fold higher risk compared to the 25–29 age group. Notably, each additional year of age after 18 increases GDM risk by approximately 7.9%, with Asian women showing a nearly double rate of increase (12.7%) compared to women of European descent (6.5%). Several studies have also reported the highest GDM prevalence among Asian women, despite their lower average age and BMI compared to White women. Findings from studies in Vietnam have demonstrated that women aged 25–34 have approximately double the risk of GDM, while those aged ≥ 35 have a threefold increased risk compared to women under 25 years of age. Another important predictor of GDM identified in our study was a history of delivering a macrosomic infant. From a pathophysiological perspective, this may reflect underlying maternal glucose intolerance in a previous pregnancy (including undiagnosed GDM), or genetic and metabolic factors that predispose to fetal overgrowth. A 2018 meta-analysis conducted in Asia showed that women with a history of macrosomia had approximately a 4.4-fold increased risk of developing GDM in subsequent pregnancies. This association has been consistently confirmed in multiple cohort studies and large-scale meta-analyses, where a previous delivery of a macrosomic infant (\u0026gt; 4,000 g) significantly increased the risk of GDM—by at least twofold—in the next pregnancy. Recent data from Vietnam also confirm that a history of delivering a macrosomic infant is associated with an increased risk of GDM (OR = 2.5; 95%CI: 1.2–5.3) [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Pre-pregnancy overweight and obesity emerged as one of the strongest risk factors for GDM in our study. Excess adiposity contributes to insulin resistance through increased fat accumulation and hormonal dysregulation, making it more difficult for the maternal body to maintain normal glucose levels during the heightened insulin demands of pregnancy [\u003cspan additionalcitationids=\"CR23\" citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e–\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]. Our findings are supported by previous data across diverse populations. A global meta-analysis showed that women who were overweight or obese before pregnancy had a 2.6-fold increased risk of developing GDM (95%CI: 1.56–4.45) compared to women with a normal BMI [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. In studies focusing on Asian populations, the risk was even higher, up to 3.27 times (95%CI: 2.81–3.80) for women with a BMI ≥ 25 compared to those with BMI \u0026lt; 25 [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]. Similarly, prior research in Vietnam found that overweight women before pregnancy had a 1.6-fold increased risk of GDM compared to women with normal weight [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. In Western populations, where obesity rates are higher and the impact of BMI is more pronounced, the risk of GDM among overweight and obese women can increase by 2 to 8 times [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. This trend clearly demonstrates that pre-pregnancy obesity places women at a significantly elevated risk of developing glucose intolerance during pregnancy. In addition to BMI, excessive gestational weight gain (GWG) is also a potential risk factor for GDM. A meta-analysis indicated that women who gained weight above the recommended levels before the typical GDM screening period (24–28 weeks) had a 40% higher risk of developing GDM (OR = 1.40; 95%CI: 1.21–1.61), regardless of their baseline BMI [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. This suggests that even women with a normal pre-pregnancy weight may face an increased GDM risk if they gain weight too rapidly or excessively during pregnancy. Multiple studies support this association. A recent large cohort study in China found that excessive weight gain during early pregnancy was associated with a 20% increased risk of GDM [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. Other studies have shown that weight gain during the first and second trimesters is more strongly associated with GDM risk than weight gain in the third trimester [\u003cspan additionalcitationids=\"CR32\" citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e–\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]. This highlights the importance of the early gestational window, particularly before 28 weeks, when excessive weight gain can significantly increase glucose intolerance risk. Clearly, excessive GWG—primarily due to abnormal fat accumulation—worsens insulin resistance, making glucose regulation more difficult, especially with early visceral fat deposition during pregnancy [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. Additionally, unhealthy dietary patterns and physical inactivity, both often associated with excessive GWG, further contribute to the elevated risk of GDM. Finally, the impact of GWG on GDM risk appears to be consistent across ethnic groups, likely due to shared physiological characteristics of pregnancy. In summary, maternal age, a history of macrosomia, elevated pre-pregnancy BMI, and excessive gestational weight gain are consistent and well-established risk factors for GDM. Pregnant women exhibiting these risk factors should be actively screened for GDM to improve early detection and intervention [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eThis study successfully developed dedicated prognostic models that allow for rapid and convenient prediction of GDM. Among these, we identified and analyzed three optimal models. Notably, Model 1 fulfilled key criteria for both simplicity and reliability. It comprises only four variables—maternal age, history of macrosomia, pre-pregnancy BMI, and gestational weight gain—all of which are routinely available in standard clinical practice. This model demonstrated superior post-test probability (0.763), the lowest Bayesian Information Criterion (BIC = − 5912), and provided acceptable discriminatory power and good predictive accuracy. The model's discrimination performance, reflected by the area under the ROC curve (AUC), was 0.74 (95%CI: 0.69–0.78). We found our model to be comparable to existing GDM prediction models worldwide. In China—a neighboring country with sociocultural characteristics similar to Vietnam—Zhang et al. reported a simple predictive model including maternal age, BMI, HbA1c, and triglycerides, with an internal AUC of 0.728 (95%CI: 0.683–0.772) [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. Gao et al. developed a GDM risk score based on six variables at the first antenatal visit (maternal age, BMI, height, family history of diabetes, systolic blood pressure, and ALT), achieving an internal validation AUC of 0.710. When four lifestyle-related variables were added, the AUC increased only slightly to 0.712 [\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]. Another model by Wu et al. aimed at early GDM prediction (before 16 weeks of gestation) included up to 15 first-trimester parameters and yielded an AUC of 0.746 [\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e]. Similarly, several other developed models have demonstrated comparable AUC values in training cohorts [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. In Europe, Benhalima et al. proposed a combined clinical model incorporating at least seven variables (including first-degree family history of diabetes, pre-pregnancy smoking, prior GDM, Asian ethnicity, maternal age, height, and pre-pregnancy BMI), achieving an internal AUC of 0.72 (95%CI: 0.69–0.76) [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]. Notably, the number of variables in these models was equal to or greater than ours, yet their discrimination performance was comparable or inferior. To enhance predictive capacity, several advanced models have incorporated early biochemical markers or utilized machine learning algorithms on large datasets. For example, when Benhalima et al. added biochemical variables, the AUC improved to 0.76 (95%CI: 0.72–0.79) [\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e]. In China, a model incorporating adiponectin and PAPP-A reached an AUC of 0.867 [\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e]. The MIDO GDM project in Mexico developed a simple artificial neural network with nine easily accessible variables, achieving an AUC of 0.85 [\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e]. A deep neural network using 73 clinical and laboratory features reached an AUC of 0.80, later simplified to a seven-variable model with an AUC of approximately 0.77 [\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e]. More complex still, a multicenter Korean study using LightGBM/XGBoost algorithms and up to 361 variables achieved an AUC of 0.80 [\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e]. While these advanced models demonstrate impressive AUC values, their complexity limits feasibility for iniial screening. More importantly, our model showed good calibration accuracy, as reflected by the Brier score. With a Brier score of 0.123, our model demonstrated well-calibrated risk predictions, a valuable strength that reflects the reliability of its estimated probabilities. In contrast, most other models emphasize discriminatory power and present calibration plots but rarely report Brier scores. The inclusion of a clear and favorable Brier score differentiates our model and highlights its potential for real-world clinical application.\u003c/p\u003e\u003cp\u003eIn addition, a common limitation of many existing models is that they are primarily evaluated using training datasets, with limited validation in independent populations. Our study’s prediction model demonstrated acceptable discriminative ability in the validation cohort, with an AUC of 0.70 (95%CI: 0.63–0.77). Compared to existing literature, our model achieved similar performance to the model by Guo and colleagues (AUC = 0.70; 95%CI: 0.68–0.72) [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e], and slightly lower than those reported by Wei et al. and Zhang et al. (AUC/C-index ranging approximately from 0.735 to 0.763) [\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e, \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e]. In a broader context, Benhalima et al. reported AUCs ranging from 0.717 to 0.769 across multiple external validation cohorts; this model is considered one of the most stable in clinical settings [\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e]. Similarly, Meertens et al. reported external validation AUCs for several models ranging from 68–75% [\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e]. However, unlike many of these models, which rely on more variables or require complex laboratory measurements, our model is parsimonious and based on easily obtainable clinical parameters. Beyond the AUC, we assessed the model's performance using calibration plots and decision curve analysis. The calibration plot showed a close alignment between the predicted probabilities and the observed outcomes, indicating that the model provides well-calibrated and reliable individual risk estimates. The DCA further demonstrated that our nomogram yielded a greater net clinical benefit than each predictor, particularly within the clinically relevant threshold probability range of 0.1 to 0.4. Other referenced models also showed good agreement between predicted and observed values in their calibration assessments [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. While these models demonstrated net clinical benefit across a broader range of threshold probabilities, their use of more numerous or complex variables may limit their feasibility for early screening, especially in primary care or low-resource settings. However, this does not necessarily imply that our model is superior, as models were developed in different populations and have not been directly cross-validated. Our model demonstrates acceptable discrimination, good calibration, and favorable net clinical benefit within a practical decision-making range. We employed Bayesian analytical principles to identify key predictors and select the optimal model for GDM prediction to achieve these outcomes. Using Bayesian Model Averaging allowed us to evaluate all possible models and derive inferences and predictions based on posterior model probabilities. This approach overcomes the limitations of traditional stepwise methods [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. As a result, our final model includes a small number of easily obtainable variables, which can be collected early in pregnancy. Importantly, these variables are well-established predictors and are commonly used as standard components in many existing GDM risk models. In summary, using robust statistical techniques such as BMA and internal bootstrap validation supports our model's simplicity, efficiency, and potential for broad clinical application in early GDM screening.\u003c/p\u003e\u003cp\u003eThe findings of this study hold important implications for the early prevention of gestational diabetes mellitus. We successfully developed a predictive nomogram to assess individualized GDM risk among pregnant women. Notably, our algorithm is applicable not only in low- and middle-income countries but also globally. Given its simplicity and ease of use, the nomogram is expected to be readily implemented in routine clinical practice. As previously discussed, the four predictors—maternal age, history of macrosomia, pre-pregnancy BMI, and gestational weight gain—can be easily obtained through patient history and routine examination. As we know, GDM increases the risk of adverse pregnancy outcomes for both mother and child. Therefore, early identification and prevention in high-risk pregnant women have the potential to reduce such complications and improve overall maternal and neonatal health. The model may be helpful in the early identification of women at high risk of developing GDM, supporting targeted monitoring or preventive interventions while reducing unnecessary procedures in low-risk individuals.\u003c/p\u003e\u003cp\u003eInterpreting our research findings requires a balanced consideration of both strengths and limitations. The study used rigorous and standardized methodology to collect a relatively adequate sample size from four independent centers. Model development incorporated internal validation via bootstrap resampling and Bayesian Model Averaging, enhancing robustness. Moreover, model performance was evaluated through discrimination, calibration plots, and decision curve analysis, ensuring comprehensive assessment. Importantly, the model was tested on an independent validation cohort, strengthening the credibility of its predictive performance. However, several limitations warrant attention. First, all study sites were located in large cities within the Mekong Delta region in southern Vietnam, which may restrict the generalizability of findings to the broader national population. Although the model showed an acceptable level of discrimination, its performance remains modest. This is partly attributable to the inherently multifactorial nature of GDM, which makes it challenging to attain substantially higher AUC values in predictive models. Furthermore, the absence of external validation in populations with distinct comorbidities or demographic profiles limits the model's applicability across diverse clinical settings. These limitations underscore the need for future research involving larger, more heterogeneous populations, including those from other geographic regions and healthcare levels.\u003c/p\u003e"},{"header":"V/ Conclusion","content":"\u003cp\u003eOur prospective multi-center cohort study showed that the independent risk factors for gestational diabetes mellitus, including maternal age, pre-pregnancy body mass index, weight gain during pregnancy, and history of macrosomia, were utilized to construct the predictive model. This model demonstrated a fairly good predictive performance and good calibration in both primary and validation cohorts, highlighting its potential as a clinical tool in identifying individuals at risk for gestational diabetes mellitus.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eAUC \u0026ndash; Area Under the Curve\u003c/p\u003e\n\u003cp\u003eBIC \u0026ndash; Bayesian Information Criterion\u003c/p\u003e\n\u003cp\u003eBMA \u0026ndash; Bayesian Model Averaging\u003c/p\u003e\n\u003cp\u003eBMI \u0026ndash; Body Mass Index\u003c/p\u003e\n\u003cp\u003eCI \u0026ndash; Confidence Interval\u003c/p\u003e\n\u003cp\u003eDCA \u0026ndash; Decision Curve Analysis\u003c/p\u003e\n\u003cp\u003eGDM \u0026ndash; Gestational Diabetes Mellitus\u003c/p\u003e\n\u003cp\u003eOGTT \u0026ndash; Oral Glucose Tolerance Test\u003c/p\u003e\n\u003cp\u003eOR \u0026ndash; Odds Ratio\u003c/p\u003e\n\u003cp\u003eROC \u0026ndash; Receiver Operating Characteristic\u003c/p\u003e\n\u003cp\u003eSD \u0026ndash; Standard Deviation\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eEthics approval\u003c/h2\u003e\u003cp\u003e All procedures performed in studies involving human participants were in accordance with the ethical standards of the Ethics Committee of University of Medicine and Pharmacy, Hue University, Vietnam (No. 1435/QĐ-ĐHYD).\u003c/p\u003e\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eConsent to participate\u003c/strong\u003e\u003cp\u003e Informed consent was obtained from all individual participants included in the study.\u003c/p\u003e\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eConsent to publish\u003c/strong\u003e\u003cp\u003eAdditional informed consent was obtained from all individual participants for whom identifying information is included in this article.\u003c/p\u003e\u003c/p\u003e\u003cp\u003e\u003ch2\u003eCompeting interests\u003c/h2\u003e\u003cp\u003eThe authors have declared that no competing interests exist.\u003c/p\u003e\u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e\u003cp\u003eThe authors have declared that there was no funding for this study.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eConception and design of the study: Nga K. Tran, Thanh N. Cao, Linh V. Pham; Acquisition of data, analysis, and interpretation of data: Nga K. Tran, Thanh N. Cao, Linh V. Pham, Tam D. Lam, Trinh A. T. Vo, Thu T. Nguyen, Nghia N. Nguyen, Dang H. Chau, Bao T. Nguyen; Final Approval: Nga K. Tran, Thanh N. Cao, Linh V. Pham, Tam D. Lam, Trinh A. T. Vo, Thu T. Nguyen, Nghia N. Nguyen, Dang H. Chau, Bao T. Nguyen.\u003c/p\u003e\u003ch2\u003eAcknowledgement\u003c/h2\u003e\u003cp\u003e We are grateful to express our sincere gratitude to the Rectorate Board of University of Medicine and Pharmacy \u0026ndash; Hue University, Can Tho Univesity of Medicine and Pharmacy, Can Tho Gynaecology and Obstetrics Hospital, An Giang Obstetric and Pediatric Hospital, Soc Trang Obstetric and Pediatric Hospital, and Ca Mau Obstetric and Pediatric Hospital for creating favorable conditions for this study to be carried out.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eAll data are available from the corresponding author by request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAmerican Diabetes Association Professional Practice Committee: 2. Diagnosis and Classification of Diabetes: Standards of Care in Diabetes\u0026mdash;2025. \u003cem\u003eDiabetes Care \u003c/em\u003e2024, 48(Supplement_1):S27-S49.\u003c/li\u003e\n\u003cli\u003eWang H, Li N, Chivese T, Werfalli M, Sun H, Yuen L, Hoegfeldt CA, Elise Powe C, Immanuel J, Karuranga S\u003cem\u003e et al\u003c/em\u003e: IDF Diabetes Atlas: Estimation of Global and Regional Gestational Diabetes Mellitus Prevalence for 2021 by International Association of Diabetes in Pregnancy Study Group\u0026apos;s Criteria. \u003cem\u003eDiabetes Res Clin Pract \u003c/em\u003e2022, 183:109050.\u003c/li\u003e\n\u003cli\u003eSweeting A, Hannah W, Backman H, Catalano P, Feghali M, Herman WH, Hivert M-F, Immanuel J, Meek C, Oppermann ML\u003cem\u003e et al\u003c/em\u003e: Epidemiology and management of gestational diabetes. \u003cem\u003eThe Lancet \u003c/em\u003e2024, 404(10448):175-192.\u003c/li\u003e\n\u003cli\u003eNguyen CL, Lee AH, Minh Pham N, Hoang Nguyen PT, Ha AVV, Khac Chu T, Van Duong D, Thi Duong H, Binns CW: Prevalence and pregnancy outcomes of gestational diabetes mellitus by different international diagnostic criteria: a prospective cohort study in Vietnam. \u003cem\u003eJ Matern Fetal Neonatal Med \u003c/em\u003e2020, 33(21):3706-3712.\u003c/li\u003e\n\u003cli\u003eLe HM, Nguyen TA, Nguyen DK, S\u0026oslash;ndergaard Linde D, Bygbjerg IC, S\u0026oslash;ndergaard J, Duc Nguyen T, Quang Nguyen B, Thi Dang NA, Nguyen XB\u003cem\u003e et al\u003c/em\u003e: Prevalence and risk factors of gestational diabetes mellitus among pregnant women in northern Vietnam: a cross-sectional study. \u003cem\u003eGlob Health Action \u003c/em\u003e2025, 18(1):2460339.\u003c/li\u003e\n\u003cli\u003eEjaz Z, Azhar Khan A, Sebghat Ullah S, Aamir Hayat M, Maqbool MA, Amin Baig A: The Effects of Gestational Diabetes on Fetus: A Surveillance Study. \u003cem\u003eCureus \u003c/em\u003e2023, 15(2):e35103.\u003c/li\u003e\n\u003cli\u003eAkinyemi OA, Weldeslase TA, Odusanya E, Akueme NT, Omokhodion OV, Fasokun ME, Makanjuola D, Fakorede M, Ogundipe T: Profiles and Outcomes of Women with Gestational Diabetes Mellitus in the United States. \u003cem\u003eCureus \u003c/em\u003e2023, 15(7):e41360.\u003c/li\u003e\n\u003cli\u003eSheiner E: Gestational Diabetes Mellitus: Long-Term Consequences for the Mother and Child Grand Challenge: How to Move on Towards Secondary Prevention? \u003cem\u003eFront Clin Diabetes Healthc \u003c/em\u003e2020, 1:546256.\u003c/li\u003e\n\u003cli\u003eStaynova R, Vasileva E, Yanachkova V: Gestational diabetes mellitus: a growing economic concern. \u003cem\u003eFolia Med (Plovdiv) \u003c/em\u003e2022, 64(5):725-732.\u003c/li\u003e\n\u003cli\u003eDall TM, Yang W, Gillespie K, Mocarski M, Byrne E, Cintina I, Beronja K, Semilla AP, Iacobucci W, Hogan PF: The Economic Burden of Elevated Blood Glucose Levels in 2017: Diagnosed and Undiagnosed Diabetes, Gestational Diabetes Mellitus, and Prediabetes. \u003cem\u003eDiabetes Care \u003c/em\u003e2019, 42(9):1661-1668.\u003c/li\u003e\n\u003cli\u003eNguyen HT, Nguyen BT, Tran AV, Nguyen TT, Ngo LH, Vo T, Nhung Thai TH, Mai LD, Tran TS, Nguyen TV\u003cem\u003e et al\u003c/em\u003e: A predictive nomogram for selective screening of asymptomatic vertebral fractures: The Vietnam Osteoporosis Study. \u003cem\u003eOsteoporos Sarcopenia \u003c/em\u003e2025, 11(1):9-14.\u003c/li\u003e\n\u003cli\u003eTran Quang B, Pham Tran P, Nguyen Thanh C, Bui Thi N, Do Dinh T, Tran Quang T, Duong Tuan L, Bui Thi Thuy N, Nguyen Anh N: High incidence of type 2 diabetes in a population with normal range body mass index and individual prediction nomogram in Vietnam. \u003cem\u003eDiabet Med \u003c/em\u003e2022, 39(2):e14680.\u003c/li\u003e\n\u003cli\u003eVo NH, Pham BV, Nguyen NN, Nguyen BT: A Predictive Nomogram for Selective Screening of Chronic Kidney Disease: A Population-Based Study. \u003cem\u003eCan J Kidney Health Dis \u003c/em\u003e2025, 12:20543581241309979.\u003c/li\u003e\n\u003cli\u003ePeduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR: A simulation study of the number of events per variable in logistic regression analysis. \u003cem\u003eJ Clin Epidemiol \u003c/em\u003e1996, 49(12):1373-1379.\u003c/li\u003e\n\u003cli\u003eAmerican Diabetes Association: 2. Classification and Diagnosis of Diabetes. \u003cem\u003eDiabetes Care \u003c/em\u003e2017, 40(Suppl 1):S11-s24.\u003c/li\u003e\n\u003cli\u003eWang D, Zhang W, Bakhai A: Comparison of Bayesian model averaging and stepwise methods for model selection in logistic regression. \u003cem\u003eStat Med \u003c/em\u003e2004, 23(22):3451-3467.\u003c/li\u003e\n\u003cli\u003eGenell A, Nemes S, Steineck G, Dickman PW: Model selection in medical research: a simulation study comparing Bayesian model averaging and stepwise regression. \u003cem\u003eBMC Med Res Methodol \u003c/em\u003e2010, 10:108.\u003c/li\u003e\n\u003cli\u003eEwout W: Clinical Prediction Models: A Practical Approach to Development, Validation, and Updating. Switzerland: Springer Cham; 2009.\u003c/li\u003e\n\u003cli\u003eHanley JA, McNeil BJ: A method of comparing the areas under receiver operating characteristic curves derived from the same cases. \u003cem\u003eRadiology \u003c/em\u003e1983, 148(3):839-843.\u003c/li\u003e\n\u003cli\u003eHarrell F: Regression Modeling Strategies: With Applications to Linear Models, Logistic and Ordinal Regression, and Survival Analysis. Switzerland: Springer Cham; 2015.\u003c/li\u003e\n\u003cli\u003eR Core Team: A language and environment for statistical computing. Vienna, Austria: Foundation for Statistical Computing; 2006.\u003c/li\u003e\n\u003cli\u003ePlows JF, Stanley JL, Baker PN, Reynolds CM, Vickers MH: The Pathophysiology of Gestational Diabetes Mellitus. \u003cem\u003eInt J Mol Sci \u003c/em\u003e2018, 19(11).\u003c/li\u003e\n\u003cli\u003eCatalano PM: Obesity, insulin resistance, and pregnancy outcome. \u003cem\u003eReproduction \u003c/em\u003e2010, 140(3):365-371.\u003c/li\u003e\n\u003cli\u003eSalameh Mohammad A, Oniya O, Chamseddine Reem S, Konje Justin C, Pan Y: Maternal Obesity, Gestational Diabetes, and Fetal Macrosomia: An Incidental or a Mechanistic Relationship? \u003cem\u003eMaternal-Fetal Medicine \u003c/em\u003e2023, 05(01):27-30.\u003c/li\u003e\n\u003cli\u003eZhang Y, Xiao CM, Zhang Y, Chen Q, Zhang XQ, Li XF, Shao RY, Gao YM: Factors Associated with Gestational Diabetes Mellitus: A Meta-Analysis. \u003cem\u003eJ Diabetes Res \u003c/em\u003e2021, 2021:6692695.\u003c/li\u003e\n\u003cli\u003eLee KW, Ching SM, Ramachandran V, Yee A, Hoo FK, Chia YC, Wan Sulaiman WA, Suppiah S, Mohamed MH, Veettil SK: Prevalence and risk factors of gestational diabetes mellitus in Asia: a systematic review and meta-analysis. \u003cem\u003eBMC Pregnancy Childbirth \u003c/em\u003e2018, 18(1):494.\u003c/li\u003e\n\u003cli\u003eChu SY, Callaghan WM, Kim SY, Schmid CH, Lau J, England LJ, Dietz PM: Maternal obesity and risk of gestational diabetes mellitus. \u003cem\u003eDiabetes Care \u003c/em\u003e2007, 30(8):2070-2076.\u003c/li\u003e\n\u003cli\u003eShin D, Song WO: Prepregnancy body mass index is an independent risk factor for gestational hypertension, gestational diabetes, preterm labor, and small- and large-for-gestational-age infants. \u003cem\u003eJ Matern Fetal Neonatal Med \u003c/em\u003e2015, 28(14):1679-1686.\u003c/li\u003e\n\u003cli\u003eBrunner S, Stecher L, Ziebarth S, Nehring I, Rifas-Shiman SL, Sommer C, Hauner H, von Kries R: Excessive gestational weight gain prior to glucose screening and the risk of gestational diabetes: a meta-analysis. \u003cem\u003eDiabetologia \u003c/em\u003e2015, 58(10):2229-2237.\u003c/li\u003e\n\u003cli\u003eYin A, Tian F, Wu X, Chen Y, Liu K, Tong J, Guan X, Zhang H, Wu L, Niu J: Excessive gestational weight gain in early pregnancy and insufficient gestational weight gain in middle pregnancy increased risk of gestational diabetes mellitus. \u003cem\u003eChin Med J (Engl) \u003c/em\u003e2022, 135(9):1057-1063.\u003c/li\u003e\n\u003cli\u003eCho EH, Hur J, Lee KJ: Early Gestational Weight Gain Rate and Adverse Pregnancy Outcomes in Korean Women. \u003cem\u003ePLoS One \u003c/em\u003e2015, 10(10):e0140376.\u003c/li\u003e\n\u003cli\u003eSommer C, M\u0026oslash;rkrid K, Jenum AK, Sletner L, Mosd\u0026oslash;l A, Birkeland KI: Weight gain, total fat gain and regional fat gain during pregnancy and the association with gestational diabetes: a population-based cohort study. \u003cem\u003eInt J Obes (Lond) \u003c/em\u003e2014, 38(1):76-81.\u003c/li\u003e\n\u003cli\u003eMacDonald SC, Bodnar LM, Himes KP, Hutcheon JA: Patterns of Gestational Weight Gain in Early Pregnancy and Risk of Gestational Diabetes Mellitus. \u003cem\u003eEpidemiology \u003c/em\u003e2017, 28(3):419-427.\u003c/li\u003e\n\u003cli\u003eZhang X, Zhao X, Huo L, Yuan N, Sun J, Du J, Nan M, Ji L: Risk prediction model of gestational diabetes mellitus based on nomogram in a Chinese population cohort study. \u003cem\u003eSci Rep \u003c/em\u003e2020, 10(1):21223.\u003c/li\u003e\n\u003cli\u003eGao S, Leng J, Liu H, Wang S, Li W, Wang Y, Hu G, Chan JCN, Yu Z, Zhu H\u003cem\u003e et al\u003c/em\u003e: Development and validation of an early pregnancy risk score for the prediction of gestational diabetes mellitus in Chinese pregnant women. \u003cem\u003eBMJ Open Diabetes Res Care \u003c/em\u003e2020, 8(1).\u003c/li\u003e\n\u003cli\u003eWu Y, Ma S, Wang Y, Chen F, Zhu F, Sun W, Shen W, Zhang J, Chen H: A risk prediction model of gestational diabetes mellitus before 16 gestational weeks in Chinese pregnant women. \u003cem\u003eDiabetes Res Clin Pract \u003c/em\u003e2021, 179:109001.\u003c/li\u003e\n\u003cli\u003eGuo F, Yang S, Zhang Y, Yang X, Zhang C, Fan J: Nomogram for prediction of gestational diabetes mellitus in urban, Chinese, pregnant women. \u003cem\u003eBMC Pregnancy Childbirth \u003c/em\u003e2020, 20(1):43.\u003c/li\u003e\n\u003cli\u003eBenhalima K, Van Crombrugge P, Moyson C, Verhaeghe J, Vandeginste S, Verlaenen H, Vercammen C, Maes T, Dufraimont E, De Block C\u003cem\u003e et al\u003c/em\u003e: Characteristics and pregnancy outcomes across gestational diabetes mellitus subtypes based on insulin resistance. \u003cem\u003eDiabetologia \u003c/em\u003e2019, 62(11):2118-2128.\u003c/li\u003e\n\u003cli\u003eBenhalima K, Van Crombrugge P, Moyson C, Verhaeghe J, Vandeginste S, Verlaenen H, Vercammen C, Maes T, Dufraimont E, De Block C\u003cem\u003e et al\u003c/em\u003e: Estimating the risk of gestational diabetes mellitus based on the 2013 WHO criteria: a prediction model based on clinical and biochemical variables in early pregnancy. \u003cem\u003eActa Diabetol \u003c/em\u003e2020, 57(6):661-671.\u003c/li\u003e\n\u003cli\u003eWang X, Sheng Y, Xiao J, Hu Y, Li L, Chen K: Combined detection of serum adiponectin and pregnancy-associated plasma protein A for early prediction of gestational diabetes mellitus. \u003cem\u003eMedicine (Baltimore) \u003c/em\u003e2024, 103(42):e40091.\u003c/li\u003e\n\u003cli\u003eGallardo-Rinc\u0026oacute;n H, R\u0026iacute;os-Blancas MJ, Ortega-Montiel J, Montoya A, Martinez-Juarez LA, Lomel\u0026iacute;n-Gasc\u0026oacute;n J, Saucedo-Mart\u0026iacute;nez R, M\u0026uacute;jica-Rosales R, Galicia-Hern\u0026aacute;ndez V, Morales-Ju\u0026aacute;rez L\u003cem\u003e et al\u003c/em\u003e: MIDO GDM: an innovative artificial intelligence-based prediction model for the development of gestational diabetes in Mexican women. \u003cem\u003eSci Rep \u003c/em\u003e2023, 13(1):6992.\u003c/li\u003e\n\u003cli\u003eWu YT, Zhang CJ, Mol BW, Kawai A, Li C, Chen L, Wang Y, Sheng JZ, Fan JX, Shi Y\u003cem\u003e et al\u003c/em\u003e: Early Prediction of Gestational Diabetes Mellitus in the Chinese Population via Advanced Machine Learning. \u003cem\u003eJ Clin Endocrinol Metab \u003c/em\u003e2021, 106(3):e1191-e1205.\u003c/li\u003e\n\u003cli\u003eKang BS, Lee SU, Hong S, Choi SK, Shin JE, Wie JH, Jo YS, Kim YH, Kil K, Chung YH\u003cem\u003e et al\u003c/em\u003e: Prediction of gestational diabetes mellitus in Asian women using machine learning algorithms. \u003cem\u003eSci Rep \u003c/em\u003e2023, 13(1):13356.\u003c/li\u003e\n\u003cli\u003eZhang J, Cao Q, Mao C, Xu J, Li Y, Mu Y, Huang G, Chen D, Deng X, Xu T\u003cem\u003e et al\u003c/em\u003e: Development and validation of a prediction model for gestational diabetes mellitus risk among women from 8 to 14 weeks of gestation in Western China. \u003cem\u003eBMC Pregnancy Childbirth \u003c/em\u003e2025, 25(1):385.\u003c/li\u003e\n\u003cli\u003eWei Y, He A, Tang C, Liu H, Li L, Yang X, Wang X, Shen F, Liu J, Li J\u003cem\u003e et al\u003c/em\u003e: Risk prediction models of gestational diabetes mellitus before 16 gestational weeks. \u003cem\u003eBMC Pregnancy Childbirth \u003c/em\u003e2022, 22(1):889.\u003c/li\u003e\n\u003cli\u003eKotzaeridi G, Bl\u0026auml;tter J, Eppel D, Rosicky I, Mittlb\u0026ouml;ck M, Yerlikaya-Schatten G, Schatten C, Husslein P, Eppel W, Huhn EA\u003cem\u003e et al\u003c/em\u003e: Performance of early risk assessment tools to predict the later development of gestational diabetes. \u003cem\u003eEur J Clin Invest \u003c/em\u003e2021, 51(12):e13630.\u003c/li\u003e\n\u003cli\u003eMeertens LJE, Scheepers HCJ, van Kuijk SMJ, Roeleveld N, Aardenburg R, van Dooren IMA, Langenveld J, Zwaan IM, Spaanderman MEA, van Gelder M\u003cem\u003e et al\u003c/em\u003e: External validation and clinical utility of prognostic prediction models for gestational diabetes mellitus: A prospective cohort study. \u003cem\u003eActa Obstet Gynecol Scand \u003c/em\u003e2020, 99(7):891-900.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-pregnancy-and-childbirth","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"prch","sideBox":"Learn more about [BMC Pregnancy and Childbirth](http://bmcpregnancychildbirth.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/prch/default.aspx","title":"BMC Pregnancy and Childbirth","twitterHandle":"@BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Gestational diabetes mellitus, risk factors, predictive model, nomogram","lastPublishedDoi":"10.21203/rs.3.rs-7009949/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7009949/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eAims\u003c/h2\u003e\u003cp\u003eGestational diabetes mellitus is commonly observed in pregnant women and is associated with an increased risk of adverse outcomes for both mother and child, not only during pregnancy but also in the long term thereafter. The present study aimed to develop a predictive nomogram for gestational diabetes mellitus in pregnant women.\u003c/p\u003e\u003ch2\u003eMaterials and methods\u003c/h2\u003e\u003cp\u003eThis multicenter prospective cohort study enrolled 1,398 pregnant women from five major obstetric hospitals in Vietnam\u0026rsquo;s Mekong Delta. GDM was diagnosed based on the 2017 American Diabetes Association criteria. Using Bayesian Model Averaging, the optimal prediction model was identified in the primary cohort (n\u0026thinsp;=\u0026thinsp;978) and used to construct a nomogram for individualized risk estimation. Model performance was validated in an independent cohort (n\u0026thinsp;=\u0026thinsp;420), with assessment of discrimination (AUC), calibration (Brier score), and clinical utility (decision curve analysis).\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e\u003cp\u003eThe prevalence of GDM was 18.0% (95% CI: 16.0\u0026ndash;20.1). The final model included maternal age (OR per year: 1.09; 95% CI: 1.06\u0026ndash;1.13), history of macrosomia (OR: 6.04; 95% CI: 2.76\u0026ndash;13.19), body mass index (OR per kg/m\u0026sup2;: 1.62; 95% CI: 1.25\u0026ndash;2.10), and weight gain during pregnancy (OR per kg: 1.12; 95% CI: 1.06\u0026ndash;1.18). The model demonstrated good discriminative ability in the primary cohort (AUC\u0026thinsp;=\u0026thinsp;0.74, Brier score\u0026thinsp;=\u0026thinsp;0.123), and acceptable performance in the validation cohort (AUC\u0026thinsp;=\u0026thinsp;0.70; 95% CI: 0.63\u0026ndash;0.77). The nomogram showed good calibration and yielded higher net benefit across a wide range of risk thresholds (0.1\u0026ndash;0.4) in decision curve analysis, indicating strong clinical utility.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e\u003cp\u003eA nomogram incorporating four routinely assessed clinical parameters offers good predictive accuracy for gestational diabetes mellitus. This model may facilitate early identification and targeted intervention for high-risk pregnant women in both resource-limited and clinical settings.\u003c/p\u003e","manuscriptTitle":"Development and Multicenter Validation of a Novel Model for Selective Screening of Gestational Diabetes Mellitus: TheVietnam Gestational Diabetes Mellitus Study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-07-14 11:35:35","doi":"10.21203/rs.3.rs-7009949/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-08-11T12:16:38+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-08-05T14:11:36+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"82069554001993726278373241152507039850","date":"2025-07-29T23:44:29+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-07-22T15:28:01+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"151742313310502022644298358335712100159","date":"2025-07-22T13:24:37+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-07-09T20:43:59+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-07-01T17:20:25+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-07-01T00:20:37+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-07-01T00:19:17+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Pregnancy and Childbirth","date":"2025-06-30T11:07:15+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-pregnancy-and-childbirth","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"prch","sideBox":"Learn more about [BMC Pregnancy and Childbirth](http://bmcpregnancychildbirth.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/prch/default.aspx","title":"BMC Pregnancy and Childbirth","twitterHandle":"@BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"c557ec34-076d-499a-944a-6dee9db15735","owner":[],"postedDate":"July 14th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-11-10T16:07:39+00:00","versionOfRecord":{"articleIdentity":"rs-7009949","link":"https://doi.org/10.1186/s12884-025-08249-w","journal":{"identity":"bmc-pregnancy-and-childbirth","isVorOnly":false,"title":"BMC Pregnancy and Childbirth"},"publishedOn":"2025-11-07 15:58:02","publishedOnDateReadable":"November 7th, 2025"},"versionCreatedAt":"2025-07-14 11:35:35","video":"","vorDoi":"10.1186/s12884-025-08249-w","vorDoiUrl":"https://doi.org/10.1186/s12884-025-08249-w","workflowStages":[]},"version":"v1","identity":"rs-7009949","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7009949","identity":"rs-7009949","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}