Nomograms for incident risk of post-partum type 2 diabetes in Chinese women with prior gestational diabetes mellitus

Predictive value of machine learning for the progression of gestational diabetes mellitus to type 2 diabetes: a systematic review and meta-analysis

BACKGROUND

This systematic review aims to explore the early predictive value of machine learning (ML) models for the progression of gestational diabetes mellitus (GDM) to type 2 diabetes mellitus (T2DM).

METHODS

A comprehensive and systematic search was conducted in Pubmed, Cochrane, Embase, and Web of Science up to July 02, 2024. The quality of the studies included was assessed. The risk of bias was assessed through the prediction model risk of bias assessment tool and a graph was drawn accordingly. The meta-analysis was performed using Stata15.0.

RESULTS

A total of 13 studies were included in the present review, involving 11,320 GDM patients and 22 ML models. The meta-analysis for ML models showed a pooled C-statistic of 0.82 (95% CI: 0.79 ~ 0.86), a pooled sensitivity of 0.76 (0.72 ~ 0.80), and a pooled specificity of 0.57 (0.50 ~ 0.65).

CONCLUSION

ML has favorable diagnostic accuracy for the progression of GDM to T2DM. This provides evidence for the development of predictive tools with broader applicability.

Risk prediction models for postpartum glucose intolerance in women with a history of gestational diabetes mellitus: a scoping review

Objective

The objective of this scoping review was to investigate the effectiveness and limitations of risk prediction models for postpartum glucose intolerance in women with gestational diabetes mellitus (GDM). The aim was to provide valuable insights for healthcare professionals in the development of robust risk prediction models.

Methods

A comprehensive literature search was conducted across multiple databases, including PubMed, EBSCO, Web of Science Core Collection, Ovid Full-Text Medical Journal Database, ProQuest, Elsevier ClinicalKey, China National Knowledge Infrastructure, China Biology Medicine, and WanFang Database, spanning from January 1990 to July 2023. To assess the quality of the included models, the Predictive Model Risk of Bias Assessment Tool (PROBAST) was employed.

Results

Fourteen relevant studies were identified and included in the final review, all focusing on model development. The discrimination ability of the included models ranged from 0.725 to 0.940, indicating satisfactory prediction accuracy. However, a notable limitation was that nine of these models (64.3%) did not provide clear guidelines on the selection of potential predictors. Furthermore, only six models (42.86%) underwent internal validation, with none undergoing external validation. A high risk of bias was observed across the included models. Logistic regression, Cox regression, and machine learning were the primary methods employed in the construction of these models.

Conclusion

The risk prediction models included in this review demonstrated favorable prediction accuracy. However, due to variations in construction methodologies, direct comparison of their performance is challenging. These models exhibited certain shortcomings, such as inadequate handling of missing data and a lack of internal and external validation, resulting in a high risk of bias. Therefore, it is recommended that these models be updated and externally validated. The development of prospective, multi-center studies is encouraged to construct predictive models with low risk of bias and high clinical applicability, ultimately guiding evidence-based clinical practice.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40200-023-01330-1.

Knowledge, attitude, and practice toward weight management among diabetic patients in Qidong City, Jiangsu Province

BACKGROUND

Weight management is an effective prevention and treatment strategy for diabetes mellitus. This study aimed to assess the knowledge, attitude, and practice (KAP) of diabetic patients towards weight management.

METHODS

Diabetic patients treated at Qidong City, Jiangsu Province, between January 2023 and June 2023 were included in this cross-sectional study. A self-designed questionnaire was used to collect their demographic characteristics and KAP toward weight management. Structural equation modeling (SEM) was employed to examine the inter-relationships among KAP scores.

RESULTS

Among a total of 503 valid questionnaires that were collected, 55.07% were filled out by men and 54.67% by those aged < 60 years. The mean scores for knowledge, attitude, and practice were 8.03 ± 3.525 (possible range: 0-13), 31.88 ± 3.524 (possible range: 10-50), and 22.24 ± 3.318 (possible range: 9-45), respectively. Pearson's correlation analysis revealed the knowledge was positively associated with attitude (r = 0.295, P < 0.001) and practice (r = 0.131, P < 0.001), and attitude was positively associated with practice (r = 0.140, P = 0.002). SEM demonstrated positive associations between knowledge and attitude (β = 0.28, P < 0.001), and attitude and practice (β = 0.09, P = 0.019). Moreover, older age was negatively associated with knowledge (β=-0.04, P = 0.001), while higher education (β = 1.220, P < 0.001), increased monthly income (β = 0.779, P < 0.001), diagnosis of fatty liver (β = 1.03, P = 0.002), and screening for excess visceral fat (β = 1.11, P = 0.002) were positively associated with knowledge.

CONCLUSION

Diabetic patients showed moderate knowledge, neutral attitudes, and inappropriate practices toward weight management. Knowledge was positively associated with attitude and practice. These findings provided valuable directions for healthcare interventions targeting improved KAP status of weight management among diabetic patients.

Precision stratification of prognostic risk factors associated with outcomes in gestational diabetes mellitus: a systematic review

BACKGROUND

The objective of this systematic review is to identify prognostic factors among women and their offspring affected by gestational diabetes mellitus (GDM), focusing on endpoints of cardiovascular disease (CVD) and type 2 diabetes (T2D) for women, and cardiometabolic profile for offspring.

METHODS

This review included studies published in English language from January 1st, 1990, through September 30th, 2021, that focused on the above outcomes of interest with respect to sociodemographic factors, lifestyle and behavioral characteristics, traditional clinical traits, and 'omics biomarkers in the mothers and offspring during the perinatal/postpartum periods and across the lifecourse. Studies that did not report associations of prognostic factors with outcomes of interest among GDM-exposed women or children were excluded.

RESULTS

Here, we identified 109 publications comprising 98 observational studies and 11 randomized-controlled trials. Findings indicate that GDM severity, maternal obesity, race/ethnicity, and unhealthy diet and physical activity levels predict T2D and CVD in women, and greater cardiometabolic risk in offspring. However, using the Diabetes Canada 2018 Clinical Practice Guidelines for studies, the level of evidence was low due to potential for confounding, reverse causation, and selection biases.

CONCLUSIONS

GDM pregnancies with greater severity, as well as those accompanied by maternal obesity, unhealthy diet, and low physical activity, as well as cases that occur among women who identify as racial/ethnic minorities are associated with worse cardiometabolic prognosis in mothers and offspring. However, given the low quality of evidence, prospective studies with detailed covariate data collection and high fidelity of follow-up are warranted.

Predictors and risk factors of short-term and long-term outcomes among women with gestational diabetes mellitus (GDM) and their offspring: Moving toward precision prognosis?

As part of the American Diabetes Association Precision Medicine in Diabetes Initiative (PMDI) - a partnership with the European Association for the Study of Diabetes (EASD) - this systematic review is part of a comprehensive evidence evaluation in support of the 2 nd International Consensus Report on Precision Diabetes Medicine. Here, we sought to synthesize evidence from empirical research papers published through September 1 st , 2021 to evaluate and identify prognostic conditions, risk factors, and biomarkers among women and children affected by gestational diabetes mellitus (GDM), focusing on clinical endpoints of cardiovascular disease (CVD) and type 2 diabetes (T2D) among women with a history of GDM; and adiposity and cardiometabolic profile among offspring exposed to GDM in utero. We identified a total of 107 observational studies and 12 randomized controlled trials testing the effect of pharmaceutical and/or lifestyle interventions. Broadly, current literature indicates that greater GDM severity, higher maternal body mass index, belonging to racial/ethnic minority group; and unhealthy lifestyle behaviors would predict a woman's risk of incident T2D and CVD, and an unfavorable cardiometabolic profile among offspring. However, the level of evidence is low (Level 4 according to the Diabetes Canada 2018 Clinical Practice Guidelines for diabetes prognosis) largely because most studies leveraged retrospective data from large registries that are vulnerable to residual confounding and reverse causation bias; and prospective cohort studies that may suffer selection and attrition bias. Moreover, for the offspring outcomes, we identified a relatively small body of literature on prognostic factors indicative of future adiposity and cardiometabolic risk. Future high-quality prospective cohort studies in diverse populations with granular data collection on prognostic factors, clinical and subclinical outcomes, high fidelity of follow-up, and appropriate analytical approaches to deal with structural biases are warranted.

Relationship between gestational body mass index change and the risk of gestational diabetes mellitus: a community-based retrospective study of 41,845 pregnant women

Background

Gestational diabetes mellitus (GDM) is associated with adverse health consequences for women and their offspring. It is associated with maternal body mass index (BMI) and may be associated with gestational weight gain (GWG). But due to the heterogeneity of diagnosis and treatment and the potential effect of GDM treatment on GWG, the association between the two has not been thoroughly clarified. Compared to body weight, BMI has the advantage that it considers height during the whole course of pregnancy. Understanding BMI changes during pregnancy may provide new evidence for the prevention of GDM.

Methods

This study investigated the BMI change of pregnant women based on a retrospective study covering all communities in Tianjin, China. According to the results of GDM screening at 24–28 weeks of gestation, pregnancies were divided into the GDM group and the non-GDM group. We compared gestational BMI change and GWG in the two groups from early pregnancy to GDM screening. GWG was evaluated according to the IOM guidelines. Logistic regression was applied to determine the significance of variables with GDM.

Results

A total of 41,845 pregnant women were included in the final analysis (GDM group, n = 4257 vs. non-GDM group, n = 37,588). BMI gain has no significant differences between the GDM and non-GDM groups at any early pregnancy BMI categories (each of 2 kg/m²), as well as weight gain (P > 0.05). Early pregnancy BMI was a risk factor for GDM (OR 1.131, 95% CI 1.122–1.139). And BMI gain was associated with a decreased risk of GDM in unadjusted univariate analysis (OR 0.895, 95% CI 0.869–0.922). After adjusting on early pregnancy BMI and other confounding factors, the effect of BMI gain was no longer significant (AOR 1.029, 95% CI 0.999–1.061), as well as weight gain (AOR 1.006, 95% CI 0.995–1.018) and GWG categories (insufficient: AOR 1.016, 95% CI 0.911–1.133; excessive: AOR 1.044, 95% CI 0.957–1.138).

Conclusions

BMI in early pregnancy was a risk factor for GDM, while BMI gain before GDM screening was not associated with the risk of GDM. Therefore, the optimal BMI in early pregnancy is the key to preventing GDM.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12884-022-04672-5.

A model for predicting gestational diabetes mellitus in early pregnancy: a prospective study in Thailand

Objective

To develop a predictive model using the risk factors of gestational diabetes mellitus (GDM) and construct a predictive nomogram for GDM risk in women during early pregnancy.

Methods

A prospective study was conducted in two tertiary hospitals among pregnant women with gestational age ≤14 weeks. Early GDM was diagnosed if an abnormal 100 g oral glucose tolerance test was detected using the Carpenter and Coustan criteria after an abnormal 50 g glucose challenge test. The factors included in the model were ACOG risk factors; maternal age; family history of hypertensive disorder in pregnancy; family history of dyslipidemia; gravida; parity; histories of preterm birth, early fetal death, abortion, stillbirth, and low birth weight; and glycated hemoglobin (HbA1c) levels. The predictive models for early GDM were analyzed using multiple logistic regression analyses. The nomograms were constructed, and their discrimination ability and predictive accuracy were tested.

Results

Of the 553 pregnant women, 54 (9.8%) were diagnosed with early GDM. In the integrated model, there was a history of GDM (adjusted odds ratio [aOR], 5.15; 95% confidence interval [CI], 1.82–14.63; P=0.004), HbA1c threshold ≥5.3% (aOR, 2.61; 95% CI, 1.44–4.74; P=0.002), and family history of dyslipidemia (aOR, 2.68; 95% CI, 1.37–5.21; P=0.005). The integrated nomogram model showed that a history of GDM had a high impact on the risk of early GDM. Its discrimination and mean absolute error were 0.76 and 0.009, respectively.

Conclusion

Application of the predictive model and nomogram will help healthcare providers investigate the probability of early GDM, especially in resource-limited countries.

Views and needs of people who at high-risk of gestational diabetes mellitus for the development of mobile health applications: A descriptive qualitative research (Preprint)

Background

Early prevention of gestational diabetes mellitus (GDM) can reduce the incidence of not only GDM, but also adverse perinatal pregnancy outcomes. Moreover, it is of great significance to prevent or reduce the occurrence of type 2 diabetes. Mobile health (mHealth) apps can help pregnant women effectively prevent GDM by providing risk prediction, lifestyle support, peer support, professional support, and other functions. Before designing mHealth apps, developers must understand the views and needs of pregnant women, and closely combine users’ needs to develop app functions, in order to better improve user experience and increase the usage rate of these apps in the future.

Objective

The objective of this study was to understand the views of the high-risk population of gestational diabetes mellitus on the development of mobile health apps and the demand for app functions, so as to provide a basis for the development of gestational diabetes mellitus prevention apps.

Methods

Fifteen pregnant women with at least one risk factor for gestational diabetes were recruited from July to September 2021, and were interviewed via a semistructured interview using the purpose sampling method. The transcribed data were analyzed by the traditional content analysis method, and themes were extracted.

Results

Respondents wanted to develop user-friendly and fully functional mobile apps for the prevention of gestational diabetes mellitus. Pregnant women's requirements for app function development include: personalized customization, accurate information support, interactive design, practical tool support, visual presentation, convenient professional support, peer support, reasonable reminder function, appropriate maternal and infant auxiliary function, and differentiated incentive function.These function settings can encourage pregnant women to improve or maintain healthy living habits during their use of the app

Conclusions

This study discusses the functional requirements of target users for gestational diabetes mellitus prevention apps, which can provide reference for the development of future applications.

Clinical predictive model for the 1-year remission probability of IgA vasculitis nephritis

PURPOSE

Early remission of Immunoglobulin A vasculitis nephritis (IgAVN) substantially affects its prognosis. In this work, a multivariate model to predict the 1-year remission probability of patients with IgAVN was developed on the basis of clinical laboratory data.

METHODS

Data of 187 patients with IgAVN confirmed by renal biopsy were retrospectively assessed. Least absolute shrinkage and selection operator regression analysis were conducted to establish a multivariate logistic regression model. A nomogram based on the multivariate logistic regression model was constructed for easy application in clinical practice. Concordance index, receiver operating characteristic (ROC) curves, decision curve analysis (DCA), and clinical impact curves (CIC) were used to evaluate the predictive accuracy and clinical value of this nomogram.

RESULTS

The predictive factors contained in the multivariate model included duration, gender, respiratory infection, arthritis, edema, estimated glomerular filtration rate, 24 h urine protein, uric acid, and renal ultrasound intensity. The area under the curves (AUC) of the nomogram in the training set and testing set were 0.814 and 0.822, respectively, indicating its good predictive ability. Moreover, the DCA curve and CIC revealed its clinical utility.

CONCLUSION

The developed multivariate predictive model combines the clinical and laboratory factors of patients with IgAVN and is useful in the individualized prediction of the 1-year remission probability aid for clinical decision-making during treatment and management of IgAVN.

Simple-to-use nomogram for evaluating the incident risk of moderate-to-severe LEAD in adults with type 2 diabetes: A cross-sectional study in a Chinese population

This study aimed to analyze the clinical characteristics of lower extremity atherosclerotic disease (LEAD) in Chinese adult type 2 diabetes (T2D) patients, and also explored the risk factors for LEAD and developed simple-to-use nomograms for LEAD and lesion degree in these patients. We retrospectively studied 4422 patients (male = 2084; female = 2338) with T2D who were ≥50. Based on lower extremity arterial ultrasound findings, we divided the patients into three groups: normal, mild, and moderate-to-severe group. The factors related to LEAD in patients with T2D were analyzed by logistic regression analysis. The risk factors for moderate-to-severe LEAD included: high HbA1c (OR = 1.07 95% CI 1.02–1.13), diabetic peripheral neuropathy (OR = 1.93 95% CI 1.57–2.37), and diabetic retinopathy (OR = 1.26 95%CI 1.01–1.57). The overall areas under the receiver operating characteristic curves for the nomograms for predicting the risks of LEAD and moderate-to-severe LEAD in adult T2D patients were 0.793 (95%CI 0.720, 0.824) and 0.736 (95%CI 0.678, 0.795), respectively. The developed nomograms are simple to use and enable preliminary visual prediction of the risk and degree of LEAD in Chinese T2D patients over 50 years. The nomograms are accurate to a certain degree and provide a clinical basis for predicting the occurrence and progression of LEAD.

Identification of hub-methylated differentially expressed genes in patients with gestational diabetes mellitus by multi-omic WGCNA basing epigenome-wide and transcriptome-wide profiling

Gestational diabetes mellitus (GDM), defined as dysglycaemia that is detected during pregnancy for the first time, has become a global health burden. GDM was found to be correlated to epigenetic changes, which would cause abnormal expression of placental genes. In the present study, we performed multi-omic weighted gene coexpression network analysis (WGCNA) to systematically identify the hub genes for GDM using both epigenome- and transcriptome-wide microarray data. Two microarray datasets (GSE70493 and GSE70494) were downloaded from the Gene Expression Omnibus (GEO) database. GEO2R was used to screen differentially expressed genes (DEGs) and differentially methylated genes (DMGs) between normal and GDM samples, separately. The results of WGCNA found that 15 modules were identified and the MEblack module had a significantly negative correlation with GDM (r = -.28, P = .03). GO enrichment analysis by BinGO of the MEblack module showed that genes were primarily enriched for the presentation of antigen processing, regulation of interferon-α production and interferon-γ-mediated signaling pathway. By comparing the DEGs, DMGs and hub genes in the coexpression network, we identified five hypermethylated, lowly expressed genes (ABLIM1, GRHL1, HLA-F, NDRG1, and SASH1) and one hypomethylated, highly expressed gene (EIF3F) as GDM-related hub DMGs. Moreover, the expression levels of ABLIM1, GRHL1, HLA-F, NDRG1, and SASH11 in the GDM patients and healthy controls were validated by a real-time quantitative polymerase chain reaction. Finally, gene set enrichment analysis showed that the biological function of cardiac muscle contraction was enriched for four GDM-related hub DMGs (ABLIM1, GRHL1, NDRG1, and SASH1). Analysis of this study revealed that dysmethylated hub genes in GDM placentas might affect the placental function and thus, take part in GDM pathogenesis and fetal cardiac development.