How common type 2 diabetes is and its risk factors in women who had diabetes during pregnancy

A thorough and systematic search was carried out across PubMed, EMBASE, Cochrane Library, and Web of Science up to March 12, 2024. Medical subject headings (MeSH) and free-text terms were used for the search. The keywords were: “diabetes mellitus, type 2”, “diabetes, gestational”, “ketosis resistant diabetes mellitus”, “non-insulin dependent diabetes”, “stable diabetes mellitus”, “MODY”, “diabetes mellitus”, “maturity onset diabetes”, “type 2 diabetes”, “adult onset diabetes”, “diabetes type ii”, “insulin independent diabetes”, “DM 2”, “T2DM”, and “pregnancy diabetes”. The details are presented in. 2

Data extraction was independently completed by two experienced reviewers. Discrepancies were addressed through discussion or consultation with a third reviewer. A standardized Microsoft Excel spreadsheet, provided by Cochrane, was used for data collection. The extracted information included the first author, publication year, country or region, study design, total sample size, prevalence of T2DM, diagnostic criteria for GDM and T2DM, and other pertinent factors. Furthermore, risk factors for T2DM identified after multifactorial logistic regression analysis, such as OR and 95% CI, were extracted. The quality and methodological rigor of all selected studies were evaluated through the Newcastle-Ottawa Scale (NOS) (14) for cohort and case-control studies, and through the Agency for Healthcare Research and Quality (AHRQ) (15) guidelines for cross-sectional studies. The NOS includes two primary components: one for cohort studies and one for case-control studies, each of which has three domains: selection, comparability, and outcome or exposure. The maximum score for each domain is nine points. The scores are categorized into three levels: low (four points or fewer), moderate (five or six points), and high (seven points or more).

The meta-analysis was performed with the help of STATA. The risk factors for T2DM were analyzed via OR values and their associated 95% CI. A fixed-effects model (FEM) was constructed for the meta-analysis when statistical heterogeneity was minimal (P > 0.10 and I² ≤ 50%); otherwise, a random-effects model (REM) was employed. In the case of observed heterogeneity, the results of the fixed-effects model and the random-effects model were compared (16). If significant discrepancies were identified, a sensitivity analysis was performed by systematically excluding each study to explore potential sources of heterogeneity. Moreover, subgroup analyses were performed to further elucidate the sources of heterogeneity. The primary outcome of this study was the risk factors for T2DM among patients with GDM. Subgroup analyses based on geographic location, study design, and other relevant variables were carried out to enhance the robustness of our findings. Risk factors demonstrating statistical significance were classified as high risk (OR≥2), moderate risk (1< OR < 2), or protective factor(OR< 1). Publication bias was assessed via funnel plots and Egger’s test. A p-value of less than 0.05 signifies statistical significance.

Our initial search yielded 22,510 articles, with 4,690 from PubMed, 6,327 from Web of Science, 9,281 from Embase, and 2,212 from Cochrane. After the exclusion of duplicates and irrelevant studies based on titles and abstracts, the full texts of the remaining 2,346 studies were reviewed. Ultimately, 46 studies were selected for inclusion. The study selection process is outlined in Figure 1.

Table 1 presents the characteristics of the included articles, along with an assessment of the quality of each study. The studies were published between 1997 and 2024. Among the studies, six (17–22) were in Europe, 15 in Asia (23–36), 21 in North America (18, 37–56), four in Oceania (57–60), and one in Africa (61). Of the ultimately selected 46 studies, 58 distinct risk factors were identified. Demographic and lifestyle factors included: age, family history of diabetes, use of progestin-only contraceptives, breastfeeding practices, higher education, ethnicity (Asian, White, Hispanic, African American), healthy dietary patterns, physical activity and sedentary behaviors, habitual iron intake, low-carbohydrate diet scores, dietary intakes, habitual alcohol consumption, and habitual coffee consumption. Pregnancy-related factors encompassed: early diagnosis of GDM, recurrence of GDM, insulin use during pregnancy, pre-pregnancy BMI, BMI during pregnancy, BMI postpartum, weight change, gestational interval, parity, waist circumference, macrosomia, neonatal birth weight, hypertension, spontaneous abortions, skin fold thickness measurements (suprailiac, tricep, subscapular), waist/hip ratio, preeclampsia, sex of the baby, and exposure to particulate matter (PM). Laboratory indicators included: HbA1c, FBG, 1-hour and 2-hour OGTT values, elevated homocysteine levels, positive autoantibodies, neonatal hypoglycemia, basal glucose clearance, OGTT total area, 30-minute insulin response during OGTT, high-density lipoprotein cholesterol, triglycerides, uric acid, circulating concentrations of branched-chain amino acids, sex hormone-binding globulin (SHBG), alanine aminotransferase (ALT), metabolomics score, and amino acid and lipid sub-scores.

Due to the absence of sufficient studies, a meta-analysis on 31 risk factors was impossible. Hence, only 26 (17–20, 22–35, 37–40, 42–44, 50–52, 57–61) of the 58 identified risk factors were meta-analyzed. The results of this analysis, as well as those from original outcome studies, are presented in Table 2. We examined the impact of these 26 risk factors on the incidence of T2DM in women with GDM. Notably, publication bias was detected for several factors, including insulin use during pregnancy, hypertension, and the 2-hour OGTT. Detailed information on publication bias can be found in Supplementary Table S2.

The quality of the 42 included cohort studies was assessed through NOS, with scores ranging from five to eight stars, indicating a relatively low risk of bias. A summary of the quality assessment is presented in. The four cross-sectional studies were evaluated using the AHRQ assessment tool, and their results also indicated reliable quality. More details are provided in. 4 4

Our study revealed that women with GDM who used progestin-only contraceptives were of advanced maternal age, or had a family history of diabetes were more likely to develop T2DM. These findings are consistent with those of Rayanagoudar et al., who also identified family history and advanced maternal age as significant risk factors for T2DM (63). They reported an RR of 1.70 for a positive family history of T2DM, which is consistent with the OR of 1.47 observed in our study. This indicates that familial factors contribute to a higher incidence of T2DM among women with a history of GDM, to some extent. The potential underlying reasons may include shared lifestyles and life philosophies within families (63). Moreover, our findings indicated a lower risk of T2DM in Caucasian women, aligning with those of Rayanagoudar et al. and You et al., who reported higher risks in Black and non-Hispanic White women after GDM (10, 63). It is important to note that, while race was not identified as a significant risk factor in our study, the observed heterogeneity in the analysis may still reflect variations in lifestyle and genetic factors across different racial groups.

Overall, patient age is a significant risk factor. As individuals age, the function of pancreatic β-cells typically declines, which directly impacts the synthesis and secretion of insulin, thereby influencing glucose regulation (64, 65). For older patients with GDM, β-cell function may have already been impaired due to aging (66, 67). The hyperglycemic stress experienced during pregnancy can further accelerate this decline, which increases their risk of developing T2DM after childbirth. Aging is also associated with heightened insulin resistance (68, 69). As individuals age, they typically experience a reduction in muscle mass and changes in visceral fat distribution, both of which contribute to systemic insulin resistance (70–72). Moreover, advancing age often coexists with inflammaging, which is a state of chronic low-grade inflammation (73, 74). Elevated levels of inflammatory markers, such as tumor necrosis factor-α (TNF-α) and interleukin-6 (IL-6), exacerbate insulin resistance and further impair β-cell function (75, 76). Lastly, although there remains ongoing debate about the link between progestin-only contraceptives and the development of diabetes, this study has found that women who use these contraceptives appear to have an increased risk of developing T2DM. Progestin-only contraceptives may contribute to this risk by inducing apoptosis in pancreatic β-cells (77), which affects blood glucose levels and disrupts glucose metabolism. This disruption can, in turn, lead to increased insulin resistance and facilitate the progression from GDM to T2DM (78, 79).

Our findings suggest that several pregnancy-related factors significantly elevate the risk of developing T2DM in women with GDM. These factors include GDM recurrence, insulin use during pregnancy, higher BMI before or after pregnancy, the delivery of macrosomic infants, and the presence of hypertension. These results are consistent with those of Rayanagoudar et al., particularly regarding BMI and insulin use during pregnancy (63). Both studies demonstrate that a high BMI substantially increases the risk of T2DM. Rayanagoudar et al. reported a progressive increase in T2DM risk with rising BMI, particularly when BMI reaches overweight or obese levels (63). This underscores the importance of managing weight before, during, and after pregnancy to prevent the progression from GDM to T2DM. Moreover, insulin use during pregnancy and GDM recurrence were identified as significant independent risk factors for T2DM. Rayanagoudar et al. found that women with GDM who required insulin therapy had a notably higher risk of developing T2DM (63). This may reflect the degree of β-cell dysfunction and the dependency on insulin during and after pregnancy. Lastly, in line with the findings of Rayanagoudar et al., our study also indicates no significant association between breastfeeding, neonatal birth weight, and the risk of T2DM in female GDM patients.

Increased insulin requirements in GDM are indicative of β-cell dysfunction. However, when β-cells cannot meet the heightened demand, GDM may develop (80). Notably, GDM patients who require exogenous insulin therapy may already have significant β-cell impairment (80). The compromised β-cell function may not fully recover postpartum, thereby raising the risk of progressing to T2DM. Additionally, a high BMI plays a crucial role in the development of insulin resistance. Obesity leads to an overproduction of inflammatory cytokines, such as TNF-α and IL-6, in adipose tissue, further exacerbating insulin resistance (81–83). Therefore, a high BMI is a significant risk factor in the progression from T2DM to GDM. Pregnancy-related complications, such as hypertension, are also important high-risk factors. Hypertension is often a marker of underlying endothelial dysfunction and systemic inflammation, both of which are closely linked to insulin resistance and the development of diabetes (84, 85). Lastly, macrosomia is another significant risk factor for the transition from GDM to T2DM, involving complex biological and physiological mechanisms. Macrosomia is typically the result of poor glycemic control during pregnancy. In GDM, insulin resistance and/or insufficient β-cell secretion lead to elevated maternal glucose levels (86, 87). These elevated glucose levels can cross the placenta, stimulate fetal growth, and lead to excessive fetal weight, or macrosomia (87). Macrosomia not only reflects the increased fetal size but also mirrors the mother’s metabolic state and insulin sensitivity (86, 87). Furthermore, the development of GDM and macrosomia is correlated with inflammation and oxidative stress (88–90). A hyperglycemic environment can promote the production of free radicals and the release of inflammatory cytokines, which in turn may damage β-cells and impair their function (88–90).

In terms of laboratory indicators, this study examined the correlation of HbA1c, FBG, as well as the one-hour and two-hour values of OGTT with the progression from GDM to T2DM. All of these indicators were proven to be significantly linked to an increased risk of developing T2DM. These parameters are crucial for monitoring glycemic control in diabetic patients, and the study’s findings further highlight their importance in predicting the progression from GDM to T2DM.

HbA1c, as a critical marker of long-term glycemic control, is particularly crucial in managing GDM. Its low heterogeneity across studies suggests that HbA1c consistently predicts T2DM risk, and can serve as a stable and reliable potential risk assessment tool. For GDM patients, persistently elevated HbA1c levels reflect prolonged hyperglycemia and may indicate further deterioration of pancreatic β-cell function (91, 92). This sustained hyperglycemic state can enhance insulin resistance, impose a greater burden on β-cells, and ultimately lead to β-cell exhaustion (91, 92). Furthermore, the analysis of FBG suggests a moderate increase in the risk of developing T2DM (OR = 1.58) and shows the importance of continued monitoring of FBG levels after pregnancy. As a tool for routine monitoring, FBG immediately reflects glycemic control and aids in the early identification of GDM patients who may develop T2DM. FBG is primarily regulated by the balance between hepatic glucose production and insulin release from pancreatic β-cells (93). For GDM patients, if β-cells fail to manage the persistent hyperglycemic stress after childbirth, their function may continue declining and lead to sustained elevations in FBG levels, which may result in the development of T2DM from GDM (94). Finally, the ORs for OGTT at one hour and two hours were 1.38 and 1.54, respectively, indicating the potential of OGTT in predicting T2DM risk. OGTT can be employed to assess insulin secretion and sensitivity by measuring an individual’s glycemic response to oral glucose (95, 96). Elevated OGTT values at one and two hours generally indicate insufficient insulin secretion or impaired insulin action (95, 96). These elevated test results physiologically reflect a diminished β-cell response to glucose and inadequate peripheral tissue response to insulin, serving as an early warning signal for T2DM development (95, 96). Therefore, systematic postpartum glycemic monitoring is essential for GDM patients, particularly those with high HbA1c and FBG levels. Regular OGTTs complement routine FBG monitoring by identifying glycemic abnormalities that may otherwise go unnoticed, facilitating early detection and timely intervention for T2DM risk.

Lastly, subgroup and sensitivity analyses were performed to delve into the prevalence and risk factors for T2DM in female GDM patients. The subgroup analysis of prevalence revealed that the incidence of T2DM among GDM patients in Asia is slightly lower compared to Europe, the Americas, and Oceania. This discrepancy may be attributed to differences in sample sizes and diagnostic criteria across regions. Additionally, the subgroup analysis of risk factors highlighted that regional variations and differences in diagnostic standards could explain the observed heterogeneity in age, insulin use during pregnancy, and FBG levels. These findings are consistent with the results of Rayanagoudar et al., who identified that follow-up duration significantly influences the risk assessment of FBG, BMI, and insulin use (63). This suggests that researchers should consider region-specific medical practices and diagnostic criteria when studying T2DM risk factors in different global regions. The sample size was identified as a source of heterogeneity for hypertension and the 1-hour OGTT, while the study design contributed to the heterogeneity observed in waist circumference and early GDM diagnosis. This highlights the critical role of study design in interpreting research findings, as variations in sampling and data collection methods can lead to biased conclusions.

Despite the aforementioned heterogeneity, sensitivity analysis confirms the stability of our findings, which are consistent with those of Rayanagoudar et al. They reported that the impact of certain key variables, such as FBG and BMI, remained significant despite variations in follow-up duration (63). This shows the reliability of the identified risk factors for the progression from GDM to T2DM, even in the presence of potential biases. Additionally, the study highlighted the potential influence of various factors on T2DM development in pregnant women, including healthful dietary patterns, physical activity, sedentary behaviors, habitual iron intake, alcohol consumption, coffee consumption, and multiple pregnancies. However, due to the limited number of original studies, the existing data are insufficient to conduct a meta-analysis on the precise effects of these factors on T2DM risk. Therefore, further research is needed to delve into these associations. It is worth noting that, in terms of the prevention of T2DM, integrative medicine research has indicated that natural remedies such as ginger and Ganoderma lucidum (lingzhi), along with their extracts, may be therapeutic and safe in modulating human metabolism (97–99). These natural agents merit further exploration as promising research foci in future studies.