Role of adiponectin in gestational diabetes mellitus: advances in mechanistic insights and early predictive potential Authors: Yuting Chen, Feiming He, Lili Zhang, Zhongying Zhang Published: 2026 ABSTRACT ABSTRACTGestational diabetes mellitus (GDM) is a common metabolic complication of pregnancy and is associated with an increased risk of short-term adverse maternal and neonatal outcomes, as well as long-term metabolic disorders in both mothers and offspring. Exacerbated insulin resistance and inadequate compensatory pancreatic β-cell function constitute the core pathophysiological basis of GDM. Adiponectin is an insulin-sensitizing adipokine that plays important roles in glucose and lipid metabolism, inflammatory regulation, and energy homoeostasis. Changes in maternal circulating adiponectin levels during pregnancy are closely associated with reduced insulin sensitivity and increased risk of GDM. This review systematically summarizes the roles of adiponectin and its distinct isoforms, particularly high-molecular-weight adiponectin, in GDM development, metabolic regulation, and early risk assessment. We also integrate potential interactions between adiponectin and placental hormones, nutrient transport, local inflammation, oxidative stress, and exosome-related mechanisms. In addition, this review discusses the key issues related to adiponectin measurement standardization, gestational timing of blood sampling, population heterogeneity, combined prediction models, and their integration into the oral glucose tolerance test (OGTT). The current findings support adiponectin as a candidate biomarker for early GDM risk stratification and combined prediction models; however, adiponectin cannot replace OGTT as a diagnostic criterion.KEYWORDS: Gestational diabetes mellitus, adiponectin, high-molecular-weight adiponectin, insulin resistance, early prediction Introduction Gestational diabetes mellitus (GDM) is a prevalent metabolic disorder that occurs during pregnancy and is a major public health concern due to its impact on maternal and foetal health. In recent years, the incidence of GDM has steadily increased, driven largely by factors such as advancing maternal age, overweight and obesity, and lifestyle changes [1]. GDM not only increases the risk of adverse pregnancy outcomes, including preeclampsia, caesarean delivery, and macrosomia, but also significantly increases the likelihood of postpartum development of type 2 diabetes mellitus (T2DM) and cardiovascular metabolic diseases in mothers, as well as long-term effects on offspring, predisposing them to obesity, impaired glucose metabolism, and an increased risk of cardiovascular disease [2–4]. Therefore, improving early detection and risk stratification of GDM is essential for optimizing short- and long-term maternal and foetal outcomes. Currently, the clinical diagnosis of GDM primarily relies on oral glucose tolerance test (OGTT) performed during the second trimester of pregnancy [5,6]. However, metabolic abnormalities are often established at this stage, limiting the window for early intervention. Consequently, identifying biomarkers that reflect metabolic dysregulation during early pregnancy has become a key research focus in the study of GDM. Insulin resistance and inadequate compensatory function of pancreatic β-cells are widely recognized as central mechanisms underlying GDM pathogenesis [7]. Adipose tissue is a vital endocrine organ that secretes various adipokines that participate in metabolic adaptation during pregnancy, thereby playing critical regulatory roles [8]. Among these adipokines, adiponectin has been extensively studied and is well known for its insulin-sensitizing, anti-inflammatory, and metabolic regulatory properties [9,10]. Adiponectin exerts its effects on multiple target tissues, including adipose tissue, liver, skeletal muscle, and pancreatic β-cells, to maintain glucose and lipid homoeostasis [11]. In recent years, clinical studies and animal models have suggested that changes in maternal circulating adiponectin levels during pregnancy are associated with reduced insulin sensitivity and increased risk of GDM. In particular, low adiponectin levels during early pregnancy may have potential value in early risk prediction [12,13]. Although previous reviews have summarized the relationship between adiponectin and GDM from the perspective of adipokines, pregnancy-related metabolic disorders, and diabetes-associated complications, several limitations remain. First, most studies have focused primarily on total adiponectin, whereas systematic comparisons of high-molecular-weight (HMW), middle-molecular-weight (MMW), and low-molecular-weight (LMW) adiponectin in terms of their structural characteristics, biological functions, inflammatory regulation, and early predictive value for GDM remain limited. Second, the interactions between adiponectin and placental hormones, placental nutrient transport, local inflammation, oxidative stress, and exosome protein cargo are not yet fully understood. Therefore, the mechanisms by which adiponectin participates in maternal – foetal metabolic regulation through the ‘maternal adipose tissue – placenta – fetus’ axis require further clarification. In addition, existing early prediction studies have mainly focused on the statistical association between adiponectin levels and GDM risk, whereas issues related to measurement methods, gestational timing of blood sampling, fasting status, cut-off values, population heterogeneity, combined prediction models, and integration with oral OGTT have not been sufficiently addressed, thereby limiting clinical translation. To address these gaps, this review systematically compares the functional differences among adiponectin isoforms and their significance in GDM risk assessment and integrates the potential links between adiponectin and placenta-related maternal – foetal interactions. Furthermore, this review discusses the potential value of adiponectin in early GDM risk stratification from the perspectives of measurement standardization, early pregnancy risk prediction, combined model construction, and clinical application boundaries. By comparing adiponectin isoforms, placenta-related mechanistic integration, and standardized early risk assessment pathways, this review aims to provide a reference for basic research, early risk identification, and clinical translation in GDM. GDM GDM is defined as hyperglycaemia that is first recognized during pregnancy but does not meet the diagnostic criteria for overt diabetes. It is one of the most prevalent metabolic complications during pregnancy. Clinical significance of GDM Pregnancy-induced hyperglycaemia (PIH) can be categorized as GDM, overt diabetes in pregnancy (ODM), and pre-GDM, with GDM accounting for approximately 83.6% of cases of PIH [14]. According to the International Diabetes Federation Diabetes Atlas, approximately 21 million neonates born annually to women aged 20–40 years are affected by PIH worldwide, 80.3% of which are attributable to GDM [15]. The global prevalence of GDM is 9–25% [16], whereas it is estimated to be 14.8% in mainland China [1]. Women with a history of GDM have an increased risk of developing T2DM, cardiovascular disease, kidney disease, hypertension, and postpartum metabolic disorders [17–21]. Early postpartum assessments indicate that 17–23% of women with previous GDM exhibit impaired glucose tolerance 4–10 weeks after delivery, whereas 5–14% develop diabetes [22]. Long-term follow-up studies suggest that the cumulative incidence of T2DM among women with a history of GDM can reach 70% over 28 years [23]. GDM is associated with numerous maternal and neonatal complications, including pre-eclampsia, pregnancy-induced hypertension, birth trauma, stillbirth, congenital malformations, macrosomia, neonatal hypoglycaemia, seizures, respiratory distress syndrome, and hyperbilirubinemia [2,24–28]. Later in life, offspring of pregnant women with GDM face an increased risk of developing obesity, cardiovascular disease, metabolic syndrome, T2DM, hypertension, and non-alcoholic fatty liver disease [3,29]. Diagnosis and management of GDM The diagnostic criteria for GDM vary across countries and regions [5,30–34]. GDM is typically screened and diagnosed after 24 weeks of gestation using the OGTT. Diagnostic thresholds are primarily based on guidelines from the American Diabetes Association, International Association of Diabetes and Pregnancy Study Groups, and World Health Organization, which recommend either a one-step or two-step testing approach [35]. In China, the diagnostic criteria for GDM include a 75-g OGTT during pregnancy: fasting plasma glucose (FPG) between 5.1–7.0 mmol/L, 1-h post-OGTT glucose of ≥10.0 mmol/L, or 2-h post-OGTT glucose between 8.5–11.1 mmol/L. GDM is diagnosed if any one of these thresholds is met, whereas an isolated elevated FPG of ≥5.1 mmol/L in early pregnancy is not sufficient for diagnosis and requires follow-up [15]. GDM is classified into two subtypes, A1 and A2. Women with A1 GDM can maintain target glucose levels through dietary modifications and physical activity, whereas those with A2 GDM require pharmacological intervention in addition to lifestyle management [36,37]. When necessary, management of GDM generally involves a combination of lifestyle modifications and pharmacotherapy. Most women with GDM achieve glycaemic control through individualized nutritional therapy and regular exercise; however, some may require insulin and/or other medications during pregnancy to maintain optimal glucose levels [38]. Hyperglycaemia during pregnancy may have long-term effects on mothers and offspring that are not necessarily resolved by the end of gestation. Therefore, it is recommended that women with GDM undergo a 75-g OGTT at 4–12 weeks postpartum to evaluate glucose metabolism. A follow-up OGTT at one year postpartum is recommended, and for women without additional risk factors, periodic 75-g OGTTs every 1–3 years are recommended to monitor metabolic status [2,39,40]. Pathogenesis of GDM The pathogenesis of GDM is complex and has not yet been fully elucidated. Current evidence suggests that the exacerbation of physiological insulin resistance during pregnancy, coupled with insufficient compensatory function of pancreatic β-cells, constitutes the core pathophysiological basis for GDM development [7,41–43]. To provide adequate glucose for foetal growth and development, all pregnant women experience a degree of physiological insulin resistance, which reduces maternal glucose utilization [41,43,44]. In women with normal glucose metabolism, FPG levels are lower than those in non-pregnant women because the foetus and placenta acquire glucose independently of insulin. Maternal insulin sensitivity is enhanced during early pregnancy but gradually declines in mid-to late pregnancy as insulin resistance increases. Following placental delivery, insulin resistance decreases rapidly [2,45,46]. In individuals with normal pancreatic function, insulin secretion is sufficient to compensate for physiological insulin resistance and maintain euglycemia [2]. Adipose tissue-derived adipokines (e.g. adiponectin and leptin), pro-inflammatory cytokines (e.g. tumour necrosis factor-α (TNF-α) and interleukin-6 (IL-6)), and placental hormones (e.g. human placental lactogen (hPL) and placental growth hormone (PGH)) all contribute to the development of physiological insulin resistance [47,48]. During normal pregnancy, maternal insulin sensitivity can decrease by approximately 50–60%, while pancreatic β-cells compensate by increasing insulin secretion to maintain glucose homoeostasis. GDM may develop when insulin resistance is further exacerbated and/or β-cell compensation is inadequate [49,50]. This manifests primarily as impaired glucose utilization in insulin-sensitive target organs, including the liver and skeletal muscles [51]. Additionally, increased maternal adiposity during pregnancy aggravates insulin resistance, and excessive gestational weight gain and prepregnancy overweight/obesity are significant risk factors for GDM [1]. The aetiology of GDM is heterogeneous. Major risk factors include prepregnancy overweight or obesity, advanced maternal age, polycystic ovary syndrome, ethnicity, family history of diabetes, and history of GDM. Other contributing factors include genetic susceptibility, dysbiosis of the gut microbiota, and vitamin D deficiency. Notably, most patients with GDM are asymptomatic, although some women may present with hyperglycaemia-related symptoms, such as polyuria, polydipsia, and fatigue; therefore, diagnosis largely relies on routine prenatal screening [2,52–54]. Overall, GDM arises from a multifactorial disruption of maternal metabolic adaptation during pregnancy, rather than a single causative factor. In addition to insulin resistance and β-cell dysfunction, various adipokines secreted by adipose tissue contribute to the regulation of glucose and lipid metabolism. Emerging evidence highlights the crucial role of adiponectin in enhancing insulin sensitivity and maintaining metabolic homoeostasis. Therefore, elucidating the biological functions of adiponectin and its metabolic regulatory effects during pregnancy is essential for a deeper understanding of the pathogenesis of GDM. Biological functions of adiponectin and its insulin-sensitizing effects Adiponectin, also known as the 30 kDa adipocyte complement-related protein (Acrp30), is an adipokine that is predominantly synthesized and secreted by adipose tissue. Since its discovery in 1995, adiponectin has been recognized as a key regulator of glucose and lipid metabolism and a critical modulator of insulin resistance [55–57]. In healthy adipose tissue, adiponectin transcription is primarily induced through the binding of peroxisome proliferator-activated receptor gamma (PPAR-γ) and CCAAT/enhancer-binding protein alpha (C/EBPα) to promoter regions [58–61], followed by translation of the full-length protein. Full-length adiponectin consists of 244 amino acids and four main domains: the N-terminal signal peptide (residues 1–18), hypervariable region (residues 18–41), collagenous domain (residues 41–107), and C-terminal globular domain (residues 107–244). Adiponectin circulates in multiple oligomeric forms, including LMW trimers, MMW hexamers, and HMW multimers [62]. Interestingly, the cleaved form, known as globular adiponectin, is generated in vivo by thrombin-mediated cleavage of the C-terminal globular domain of full-length adiponectin (Figure 1). In adipocytes, the transcription of adiponectin is primarily induced by the binding of PPAR-γ and C/EBPα to a promoter region. After transcription, the full-length adiponectin protein, comprising 244 amino acids, is translated and contains four major domains: N-terminal signal sequence (1–18), hypervariable region (18–41), collagenous domain (41–107), and C-terminal globular domain (107–244). Full-length adiponectin circulates in the body in different molecular forms, including LMW trimers, MMW hexamers, and HMW multimers. Globular adiponectin is generated by thrombin-mediated cleavage of the C-terminal globular domain of full-length adiponectin. Adiponectin predominantly circulates in plasma as multimeric complexes, with concentrations typically ranging from approximately 5 to 30 μg/mL, accounting for about 0.01% of total plasma protein [63]. Clearance kinetic studies have shown that different adiponectin multimers exhibit distinct in vivo clearance rates, with HMW adiponectin being the slowest cleared form [64]. Adiponectin is mainly taken up and cleared by the liver and can also be excreted through the kidneys, which may be related to differences in molecular size and tissue transport properties among its multimeric forms. In terms of metabolic regulation, HMW adiponectin is generally considered the principal functional form responsible for the metabolic protective effects of adiponectin. Multiple population-based studies have demonstrated that circulating HMW adiponectin levels and the HMW-to-total adiponectin ratio are closely associated with insulin sensitivity, metabolic syndrome, and T2DM [65–67]. By contrast, the independent biological function of MMW adiponectin remains unclear, although it may be more responsive to acute physiological stimuli. For example, one study found that a single session of high-intensity interval training significantly increased MMW adiponectin levels, whereas HMW and LMW adiponectin levels remained unchanged, suggesting the degree of specificity of MMW adiponectin in the acute exercise response [68]. LMW adiponectin may have greater tissue accessibility and a stronger tendency for renal filtration. One study reported that urinary adiponectin exists predominantly in the LMW form, suggesting that it may leak through or be filtered by the kidneys more readily than higher-order multimers [69]. Different adiponectin multimers appear to exert distinct effects on inflammation regulation. Neumeier et al. [70] showed that adiponectin multimers share some common effects but also display multimer-specific actions, with LMW adiponectin showing a more evident anti-inflammatory profile. Schober et al. [71] further reported that HMW adiponectin enhances IL-6 release from monocytes, whereas LMW adiponectin reduces IL-6 release, suggesting that decreased LMW adiponectin in obesity or T2DM may weaken its anti-inflammatory regulatory effects. Overall, the functional differences among adiponectin isoforms are not determined simply by molecular size but are closely related to their multimeric structure, clearance kinetics, tissue distribution, and target-cell signalling responses. Adiponectin exerts its biological effects primarily through its classical receptors, adiponectin receptor 1 (AdipoR1) and adiponectin receptor 2 (AdipoR2), which AdipoR1 predominantly expressed in skeletal muscle. AdipoR2 is highly expressed in the liver. Upon binding to its receptors, adiponectin activates downstream signalling pathways, including AMP-activated protein kinase (AMPK) and peroxisome proliferator-activated receptor alpha (PPAR-α), thereby promoting skeletal muscle glucose uptake, mitochondrial biogenesis, and fatty acid oxidation, while inhibiting hepatic gluconeogenesis and enhancing fatty acid oxidation. Collectively, these effects improve insulin sensitivity and maintain glucose and lipid homoeostasis (Figure 2) [72–75]. As a widely recognized insulin-sensitizing adipokine, adiponectin not only enhances systemic insulin responsiveness through the regulation of energy metabolism in skeletal muscle, liver, and adipose tissue but also mitigates obesity-associated chronic low-grade inflammation via anti-inflammatory and antioxidative effects, further supporting its role in improving insulin sensitivity [9]. In studies on prediabetes and T2DM, hypoadiponectinemia was closely associated with the progression of insulin resistance, whereas increasing adiponectin levels or enhancing its signalling may help delay the transition from metabolic dysfunction to an overt diabetes transition [76]. Adiponectin also contributes to the regulation of maternal insulin sensitivity during pregnancy. Studies indicate that circulating adiponectin levels decrease as pregnancy progresses, corresponding to a reduction in insulin sensitivity that is more pronounced in women who develop GDM, suggesting that insufficient adiponectin levels may exacerbate gestational insulin resistance [51]. Therefore, adiponectin not only serves as an important biomarker of insulin sensitivity but also as a potential therapeutic target for the management of insulin resistance and associated metabolic disorders [9,72]. Structure, receptors, and function of adiponectin in skeletal muscle and the liver. gAd, LMW, MMW, and HMW bind to AdipoR1 and AdipoR2, with AdipoR1 highly expressed in skeletal muscle and AdipoR2 highly expressed in the liver. Binding of adiponectin to its receptors activates downstream signalling pathways, including AMPK and PPAR-α, thereby promoting glucose uptake, mitochondrial biogenesis, and fatty acid oxidation in skeletal muscle while inhibiting hepatic gluconeogenesis and enhancing hepatic fatty acid oxidation. These effects collectively improve insulin sensitivity and maintain glucose and lipid homoeostasis. Adiponectin and GDM Given the critical role of adiponectin in regulating insulin sensitivity and maintaining glucose and lipid homoeostasis, it has attracted increasing attention due to its involvement in metabolic adaptation during pregnancy and GDM pathogenesis. Emerging evidence indicates that adiponectin not only participates in the regulation of multiple maternal organs during pregnancy, but is also closely associated with GDM risk, disease severity, and early predictive potential. Physiological roles of adiponectin during pregnancy During pregnancy, the maternal body undergoes a series of adaptations in glucose and lipid metabolism, as well as in pancreatic islet cell function, to meet the energy demands of foetal growth and development. Adiponectin may participate in the regulation of maternal energy metabolism, lipid homoeostasis, and insulin sensitivity by acting on multiple target organs, including adipose tissue, liver, skeletal muscle, and pancreas (Figure 3). It should be emphasized that the mechanistic basis for the organ-specific actions of adiponectin is mainly derived from mouse models and basic experimental studies, whereas human data are largely based on longitudinal observational or case-control studies. Therefore, findings from animal and human studies should be interpreted separately. Role of adiponectin in pregnancy. In mice, normal late pregnancy is accompanied by reduced circulating adiponectin levels and decreased adipose tissue Adipoq mRNA expression. Adiponectin gene deletion further enhances adipose tissue lipolysis, increases hepatic glucose and triglyceride production, and leads to impaired glucose tolerance and hyperlipidaemia in late pregnancy. Another mouse study showed that adiponectin deficiency induced marked hepatic steatosis, whereas adiponectin supplementation in late pregnancy attenuated hepatic steatosis and restored normal glucose homoeostasis [13,77,78]. However, Qiao et al. [77] found that adiponectin deficiency did not further reduce insulin-stimulated glucose clearance in peripheral tissues, suggesting that the observed glucose metabolic abnormalities may be more closely related to increased hepatic glucose output, enhanced lipolysis, and inadequate pancreatic islet compensation, rather than simply attributable to peripheral insulin resistance. Mouse studies have also suggested that adiponectin contributes to maternal β-cell expansion by promoting placental lactogen expression. Although this upstream effect has been supported in vitro using human placental explants and trophoblast cells, the in vivo basis for adiponectin-mediated maternal β-cell expansion during pregnancy remains mainly derived from mouse models [79]. In addition, the ability of adiponectin to promote glucose uptake and fatty acid oxidation in skeletal muscle has been supported by basic mechanistic studies [80,81], but direct organ-level evidence in human pregnancy remains limited. Longitudinal human studies have shown consistency with mouse studies regarding the overall direction of adiponectin changes during pregnancy. Catalano et al. [82] followed normal-weight women before pregnancy and during early and late pregnancy and found that total adiponectin and HMW adiponectin levels were reduced in late pregnancy compared with those in the prepregnancy state, accompanied by decreased adipose tissue ADIPOQ mRNA expression. These changes were mainly associated with reduced insulin-mediated glucose disposal, whereas their relationship with endogenous glucose production or lipid metabolism were less evident. In human GDM studies, maternal circulating adiponectin levels and adiponectin expression in the subcutaneous and visceral adipose tissues have been reported to decrease, and some differences remain significant after adjusting for body mass index (BMI) [83]. Overall, both mouse and human studies suggest that dynamic changes in adiponectin levels during pregnancy are closely related to maternal metabolic adaptation; however, the timing, magnitude, and specific mechanisms of these changes are not fully consistent between species. Mouse studies commonly rely on complete gene deletion or exogenous adiponectin supplementation, producing effects that are much stronger than the relative hypoadiponectinemia typically observed in humans with GDM. In contrast, human findings are influenced by prepregnancy BMI, fat distribution, ethnicity, gestational age, and treatment-related factors. Thus, animal experiments provide causal and mechanistic support for the involvement of adiponectin in maternal lipolysis, hepatic lipid metabolism, and pancreatic islet-cell adaptation; however, the magnitude of these effects and the specific pathways involved require further validation in human longitudinal and interventional studies. Maternal circulating adiponectin levels and GDM During normal pregnancy, maternal circulating adiponectin levels exhibit dynamic fluctuations; however, this decline is often more pronounced in women who develop GDM. Numerous studies have demonstrated that hypoadiponectinemia is closely associated with the onset of GDM, exacerbated insulin resistance, and adverse perinatal outcomes. Evidence indicates that maternal adiponectin levels decrease during pregnancy relative to the non-pregnant state, and this reduction is more pronounced in women with GDM [12,84–86]. Wang et al. [85] reported that serum adiponectin levels were significantly lower in patients with GDM than in normoglycemic pregnant women and that adiponectin levels were closely correlated with the degree of insulin resistance, suggesting a potential role for hypoadiponectinemia in GDM pathogenesis. Furthermore, Muntean et al. [86] observed that reduced adiponectin levels in GDM were often accompanied by elevated insulin levels and homoeostasis model assessment of insulin resistance scores and were associated with certain adverse perinatal outcomes, indicating that adiponectin abnormalities may not only reflect maternal metabolic imbalance but also contribute to maternal – foetal complications. Recent studies have further confirmed a strong association between low adiponectin levels and an increased risk of developing GDM. Gao et al. [12] conducted a meta-analysis of 28 studies involving 12,256 pregnant women and found that circulating adiponectin levels were significantly lower in the GDM group than in normoglycemic controls. Importantly, high adiponectin levels were associated with a reduced risk of GDM, highlighting the potential predictive value of adiponectin. Similarly, Gruber et al. [51] noted that adiponectin levels decline during pregnancy in parallel with reduced maternal insulin sensitivity. This hypoadiponectinemic state is more pronounced in women with GDM, supporting adiponectin as a candidate biomarker for early GDM risk assessment. At the maternal – foetal interface, interactions between placental hormones and adiponectin may represent an important link connecting maternal metabolic adaptation, placental function, and the foetal metabolic environment. hPL and PGH are key placenta-derived metabolic hormones produced during pregnancy. Their circulating levels increase progressively with advancing gestation and contribute to maternal energy metabolic remodelling by reducing maternal insulin sensitivity, promoting lipolysis, and increasing free fatty acid availability [87–89]. These metabolic adaptations may influence the endocrine function of maternal adipose tissue, thereby indirectly contributing to the decline in adiponectin levels and the development of hypoadiponectinemia during pregnancy [87,90]. Maternal circulating adiponectin, particularly HMW adiponectin, is unlikely to cross the placenta in physiologically meaningful amounts. Therefore, its effects on foetal metabolism are more likely to be mediated indirectly through modulation of the local placental microenvironment [90,91]. Adiponectin receptors AdipoR1 and AdipoR2 mediate downstream metabolic signalling pathways, including AMPK and PPARα [92], and adiponectin and its receptors are also expressed in the placenta and trophoblast cells [93]. Accordingly, adiponectin may act on trophoblast cells and vascular endothelial-related pathways to regulate placental insulin signalling, mTOR activity, and the transport of nutrients, such as glucose, amino acids, and fatty acids [90,91,94]. In addition, adiponectin may modulate AMPK, nuclear factor kappa B, mitogen-activated protein kinase, and oxidative stress-related signalling pathways, thereby influencing the expression of pro-inflammatory cytokines such as TNF-α and IL-6 and reducing reactive oxygen species generation and oxidative injury. These effects may help to maintain placental endothelial function and metabolic homoeostasis [81,95]. Therefore, although adiponectin does not primarily act on the foetus through direct transplacental transfer, it may indirectly regulate the intrauterine nutritional environment and long-term foetal metabolic health by influencing maternal metabolic status, placental inflammation, oxidative stress, and nutrient transport [81,90,91,94,95]. In addition to reduced levels of circulating free adiponectin, adiponectin abnormalities in GDM may also be reflected in the protein composition of peripheral blood exosomes. Using iTRAQ-based proteomic analysis, Lin et al. [96] found that adiponectin protein abundance was significantly decreased in total plasma exosomes from women with GDM at 24–28 weeks of gestation (fold change = 0.54, p = 0.047), and adiponectin was identified as a key protein related to T2DM and the PPAR signalling pathway. However, at 12–16 weeks of gestation, the reduction in adiponectin levels did not reach statistical significance (fold change = 0.51, p = 0.278). Notably, adiponectin was not identified as a differentially expressed protein in placenta-derived exosomes in this study. Therefore, these findings more likely reflect altered adiponectin cargo in total circulating exosomes; however, its tissue origin and functional significance require further investigation. This study suggests that GDM-related adiponectin abnormalities may involve not only reduced serum or plasma free adiponectin levels, but also changes in exosomal protein cargo, providing a new perspective for understanding the different forms of adiponectin alterations in GDM. In addition to its association with disease occurrence, adiponectin may also reflect treatment response and disease progression in GDM. Wei et al. [97] reported that adiponectin levels increased after a clinical intervention in patients with GDM, suggesting its potential as a marker of metabolic improvement. Similarly, Muntean et al. [98] found that adiponectin levels at delivery were significantly lower in women with GDM than in controls, and that low adiponectin levels were associated with a high incidence of adverse outcomes, such as pregnancy-induced hypertension and macrosomia, indicating that hypoadiponectinemia may persist throughout GDM progression and contribute to perinatal complications. Animal studies have provided further evidence supporting the role of adiponectin in the pathogenesis of GDM. Gruber et al. [13] compared pregnant adiponectin-knockout mice with wild-type controls at 3 weeks of gestation and found that adiponectin deficiency led to fasting hyperglycaemia, impaired glucose and pyruvate tolerance, marked hepatic steatosis, and elevated hepatic triglycerides. Supplementation with adiponectin at 2 weeks of gestation substantially ameliorated these abnormalities, suggesting that adiponectin deficiency can induce a metabolic phenotype resembling GDM, and may contribute to disease development via dysregulated gluconeogenesis and hepatic lipid accumulation. Additionally, in a streptozotocin-induced GDM rat model, Cheng et al. [99] demonstrated that adiponectin alleviated GDM-associated metabolic disturbances by upregulating sex hormone-binding globulin expression and improving insulin signalling, highlighting its potential as a therapeutic target for GDM. In summary, reduced maternal circulating adiponectin levels are closely associated with the onset of GDM, exacerbated insulin resistance, and adverse perinatal outcomes. Adiponectin not only serves as a key biomarker reflecting maternal metabolic status during pregnancy but also holds promise as a potential tool for early GDM screening, risk stratification, and targeted intervention. Potential value of adiponectin in early prediction of GDM Currently, GDM screening and diagnosis are typically conducted during the second trimester of pregnancy using the OGTT. The interval between diagnosis and delivery is relatively short, which limits the window for effective intervention. Therefore, early identification of high-risk pregnant women is critical for implementing timely lifestyle modifications and clinical interventions to improve maternal and neonatal outcomes. Adiponectin, which reflects maternal insulin sensitivity and glucose – lipid metabolic status in early pregnancy, has garnered considerable attention as a potential early predictive biomarker for GDM. Clinical studies suggest that lower adiponectin levels during early pregnancy are associated with an increased risk of subsequent GDM. Yang et al. [100] reported that, after adjusting for maternal age and family history of diabetes, HMW adiponectin levels at 11–14 weeks of gestation were significantly lower in women who later developed GDM than in normoglycemic controls. Hang et al. [101] reported that serum adiponectin levels measured in early pregnancy (<13 weeks of gestation) predicted subsequent GDM, with a receiver operating characteristic (ROC) curve analysis showing an area under the curve (AUC)of 0.81 (95% CI: 0.72–0.91). At the optimal cut-off value of 8.41 μg/mL, the sensitivity and specificity were 71.43% and 80.95%, respectively, indicating a moderate predictive performance. Similarly, Fang et al. [102] showed that adiponectin measured at 7–13 weeks of gestation had a good predictive performance for GDM, with an ROC-AUC of 0.962 (95% CI: 0.918–1.000). When 4.24 μg/mL was used as the cut-off value, the sensitivity and specificity were 90.0% and 96.70%, respectively. Bai et al. [103] developed an early GDM risk prediction model and found that women who subsequently developed GDM had lower adiponectin levels than those with normal pregnancies at 7–12 weeks of gestation. In that study, adiponectin predicted subsequent GDM with an ROC-AUC of 0.723 and a cut-off value of 13.38 mg/L, suggesting that adiponectin may have potential value for early GDM risk identification. Madhu et al. [104] reported an optimal cut-off value of 9.10 μg/mL for adiponectin at 11–13 weeks, corresponding to a sensitivity of 100% and specificity of 95.6%. Collectively, these findings suggest that early measurement of maternal serum adiponectin levels may provide valuable predictive information for GDM (Table 1). Table 1.Predictive performance of maternal circulating adiponectin for early risk assessment of gestational diabetes mellitus.Author (year) (ref.)Gestational weeksROC-AUC95% CICut-off valueSensitivitySpecificityYang et al. (2024) [100]11–14 weeks–––––Hang et al. (2022) [101]<13 weeks0.810.720–0.9108.41 ug/mL71.43%80.95%Fang et al. (2025) [102]7–13 weeks0.9620.918–1.0004.24 ug/mL90%96.70%Bai et al. (2025) [103]7–12 weeks0.723–13.38 ug/mL––Madhu et al. (2019) [104]11–13 weeks––9.10 ug/m100%95.6% Predictive performance of maternal circulating adiponectin for early risk assessment of gestational diabetes mellitus. Among the different adiponectin isoforms, HMW adiponectin currently has the most substantial research basis and clearest clinical relevance in GDM risk assessment and early prediction. First, HMW adiponectin is closely related to insulin sensitivity and glucose – lipid metabolic homoeostasis [65–67], which is consistent with the core pathophysiology of GDM characterized by exacerbated insulin resistance and inadequate β-cell compensation. Second, lower HMW adiponectin levels can be observed before or during early pregnancy and are associated with an increased risk of subsequent GDM [105]. Third, longitudinal studies in multiethnic cohorts have shown that incorporating HMW adiponectin into first-trimester risk models may improve the predictive performance of GDM to some extent [106]. In addition, HMW adiponectin is negatively correlated with BMI and the homoeostatic model assessment of insulin resistance, suggesting that it may reflect obesity-related insulin resistance and adipose tissue dysfunction. Dynamic postpartum changes may also provide clues for assessing long-term glucose metabolic risk in women with a history of GDM [107]. In contrast, the independent predictive value of MMW and LMW adiponectin in GDM remains insufficiently supported by current research. Therefore, HMW adiponectin may be a more suitable candidate biomarker for early GDM risk stratification and combined prediction models, although it cannot replace the OGTT as a diagnostic criterion for GDM. However, the ROC-AUC values, cut-off points, and predictive performance reported across studies vary considerably, suggesting that adiponectin as a single predictive marker is still affected by population heterogeneity and potential confounding factors. Ethnic background, prepregnancy BMI, and the degree of obesity can all influence adiponectin levels and their predictive performance for GDM. Obesity itself may reduce the total adiponectin and HMW adiponectin levels, leading to overlapping values between women with and without GDM, thereby weakening the specificity and discriminative ability of fixed cut-off points [105,107,108]. Although the association between low adiponectin levels and GDM risk may persist after adjusting for BMI and ethnicity [105,106], substantial variation remains in the optimal cut-off and ROC-AUC values reported across populations, and unified ethnicity- or BMI-specific reference ranges have not yet been established [104,109,110]. Simultaneously, the standardization of adiponectin measurement and risk assessment workflows is essential for its clinical translation. Maternal circulating adiponectin levels can be measured in serum or plasma using immunoassays. Total adiponectin is commonly detected using enzyme-linked immunosorbent assay (ELISA), whereas HMW adiponectin can be measured using isoform-specific ELISA or chemiluminescent immunoassays. When simultaneous differentiation of HMW, MMW, and LMW adiponectin is required, research-based techniques, such as native electrophoresis combined with western blotting or gel filtration, are often used [111–113]. As different assay platforms vary in antibodies, calibrators, sample processing procedures, and reporting units, the absolute concentrations and cut-off values reported across studies should not be directly interchanged [111,112]. No mandatory fasting requirement for adiponectin measurement is currently available; however, to reduce pre-analytical variability, blood sampling time and feeding status should be standardized as much as possible. If adiponectin levels are assessed together with fasting glucose, insulin, or lipid profiles, morning fasting blood samples may be more appropriate. For early GDM risk assessment, adiponectin can be measured at the first antenatal visit or during first-trimester screening at 10–13 + 6 weeks of gestation, thereby preserving a longer window for risk intervention [106,110]. Repeated measurements during pregnancy may be useful for studying the dynamic trajectory of adiponectin; however, there is currently insufficient clinical evidence to support routine repeated testing in standard antenatal care [106,114]. Therefore, adiponectin is better used as a continuous variable in combined risk models for auxiliary risk stratification rather than as a single marker to replace OGTT or to directly diagnose GDM [106,110,114]. In view of the limited predictive performance of adiponectin alone and considerable heterogeneity among study populations, recent research has increasingly focused on combined biomarker-based models for GDM risk prediction. Inthavong et al. [115] reviewed evidence indicating that adipokines, inflammatory markers, insulin resistance-related indices, and other biochemical biomarkers have the potential to predict GDM. However, no single biomarker can replace existing diagnostic standards, highlighting the potential need for combined biomarker strategies to improve predictive accuracy. Supporting this, Wang et al. [116] reported that combined assessment of maternal serum adiponectin and pregnancy-associated plasma protein A (PAPP-A) predicted GDM with an ROC-AUC of 0.867, with both sensitivity and specificity higher than those of either marker alone. These findings suggest that the combination of adiponectin and PAPP-A may improve early identification of GDM risk. PAPP-A is a placenta-derived glycoprotein that regulates the bioavailability of insulin-like growth factors and participates in placental development, foetal growth, and maternal glucose metabolic regulation [117]. Reduced PAPP-A levels during early pregnancy may reflect impaired placental function and have been associated with an increased risk of GDM [118,119]. From a mechanistic perspective, adiponectin and PAPP-A may represent complementary biological domains that reflect maternal adipose tissue function, insulin sensitivity, and placenta-related metabolic alterations. Therefore, their combined use in early GDM risk assessment has a plausible biological rationale. Based on the current findings, a candidate first-trimester risk stratification model incorporating HMW adiponectin, PAPP-A, and maternal clinical risk factors may be further developed for predicting GDM [106,116,120–122]. This model can be applied at 11–13 + 6 weeks of gestation. Maternal age, prepregnancy BMI, history of GDM, family history of diabetes, previous macrosomia, parity, polycystic ovary syndrome, and ethnic or population background can be used to estimate baseline risk. Quantitative HMW adiponectin levels and PAPP-A multiples of the median (MoM), adjusted for gestational age and relevant maternal factors, could then be incorporated to improve the completeness and accuracy of first-trimester risk stratification. This combination may provide more comprehensive risk information than a single marker by simultaneously reflecting maternal insulin resistance, adipose tissue dysfunction, and placenta-related metabolic changes. It should be emphasized that no study has externally validated the specific combination of HMW adiponectin, PAPP-A, and the clinical risk factors. The closest available evidence is derived from studies evaluating the combined detection of total adiponectin and PAPP-A levels [116]. Additionally, the predictive performance of PAPP-A alone for GDM is limited, with substantial heterogeneity across studies [119,123]. The incremental value of HMW adiponectin beyond conventional clinical risk factors may also be influenced by differences in GDM diagnostic criteria, study populations, and assay platforms [120]. Therefore, at the current stage, this model should be regarded as a risk stratification tool for identifying women who may benefit from early confirmatory glucose testing, rather than as a substitute for OGTT or an independent diagnostic tool for GDM [114,123]. Its clinical translation depends on several key issues, including the standardization of HMW adiponectin assays, cross-platform calibration of the PAPP-A MoM, external validation across different ethnic and BMI subgroups, determination of clinically meaningful risk thresholds and cost-effectiveness, and prospective validation of whether model-guided early screening and intervention can improve maternal and neonatal outcomes [111,114,124,125]. Discussion The development and progression of GDM are closely linked to impaired metabolic adaptation during pregnancy, with exacerbated insulin resistance and insufficient compensatory pancreatic β-cell function representing its core pathophysiological basis. As an important insulin-sensitizing adipokine, adiponectin participates in glucose and lipid metabolism, energy homoeostasis, and inflammatory regulation by acting on multiple target organs, including adipose tissue, liver, skeletal muscle, and pancreas. Overall, current studies suggest that reduced maternal circulating adiponectin levels, particularly decreased HMW adiponectin levels, are associated with GDM occurrence, increased insulin resistance, and adverse pregnancy outcomes. However, mechanistic insights derived from animal experiments are not equivalent to the clinical associations observed in human studies. Moreover, differences in study populations, gestational timing of blood sampling, adiponectin isoforms, assay methods, and GDM diagnostic criteria limit the direct comparability and clinical extrapolation of the existing findings. From a clinical perspective, adiponectin is considered more appropriate as an auxiliary marker for early GDM risk stratification than as an independent diagnostic biomarker. Measurement of adiponectin, especially HMW adiponectin, in early pregnancy may help identify women with greater insulin resistance or a higher subsequent risk of GDM, thereby providing a basis for earlier confirmatory glucose testing, lifestyle counselling, and individualized follow-up. Combining adiponectin levels with prepregnancy BMI, maternal age, history of GDM, family history of diabetes, PAPP-A, and other metabolic indicators may provide more comprehensive risk information than a single biomarker. However, adiponectin levels are influenced by BMI, ethnicity, gestational age, inflammatory status, and assay platform, and no unified cut-off value has yet been established. Therefore, low adiponectin levels should not be used directly as a diagnostic criterion for GDM or as an indication for initiating treatment, and normal adiponectin levels cannot exclude GDM. The OGTT performed at 24–28 weeks of gestation remains the primary diagnostic approach for GDM in current clinical practice. Future studies should address several specific issues. First, large-scale, multicenter, prospective longitudinal cohorts with standardized sampling time points before pregnancy or at 10–13 + 6 weeks of gestation, mid pregnancy, late pregnancy, and the postpartum period are required. These studies should dynamically evaluate the trajectories of total adiponectin and its HMW, MMW, and LMW isoforms and assess their relationships with GDM occurrence, pregnancy outcomes, and postpartum glucose metabolic status. Second, sample type, fasting status, blood collection time, sample storage conditions, assay platforms, calibrators, and quality control procedures should be standardized to reduce preanalytical and analytical variability across studies. Third, prediction models incorporating HMW adiponectin, clinical risk factors, PAPP-A, or other metabolic and inflammatory markers should be validated in populations stratified by BMI, ethnicity, and GDM diagnostic criteria. Model performance should be systematically reported using ROC-AUC, sensitivity, specificity, calibration performance, decision curve analysis, net benefit, and external validation, rather than relying solely on optimal cut-off values derived from single-centre studies. In addition, mechanistic studies should further integrate animal pregnancy models, human placental explants, trophoblast cells, and organoid systems to clarify the interactions between adiponectin and placental hormones, nutrient transport, inflammatory responses, oxidative stress, and exosomal protein cargo, while clearly defining the translational boundaries between animal findings and human pregnancy. Finally, model-guided clinical impact studies are needed to determine whether adiponectin-based early risk stratification can accurately identify high-risk women earlier and improve clinically relevant outcomes, including glycaemic control, macrosomia, gestational hypertension, caesarean delivery, and neonatal complications. Only after assay standardization, external validation across populations, and demonstration of clinical benefits, can adiponectin be translated from a research biomarker into a practical tool for early GDM risk management. Funding Statement This work was supported by the Natural Science Foundation of Xiamen under Grant [3502Z202374078]. Disclosure statement No potential conflict of interest was reported by the author(s). Data availability statement Data sharing is not applicable to this article as no data were created or analysed in this study. References Associated Data Data Availability Statement Data sharing is not applicable to this article as no data were created or analysed in this study.