Skip to main content
Download PDF

Complex interplay between type 2 diabetes mellitus and pancreatic cancer: insights from observational and mendelian randomization analyses

BMC Cancer volume 25, Article number: 556 (2025) Cite this article

Abstract

Background

To investigate the causal relationship between type 2 diabetes mellitus (T2DM), pancreatic cancer (PC) risk and identify the mediating effects of various risk factors on that relationship.

Methods

581 PC patients and 582 healthy controls who visited our center from January 2013 to December 2023 were included in this retrospective study. Multivariable logistic regression was performed to evaluate the association between T2DM and PC through odds ratios (ORs) and 95% confidence intervals (CIs). Mendelian randomization (MR) studies were then conducted to explore the causal relationship between T2DM and PC, and causal mediation analysis (CMA) to examine the mediating role of common risk factors.

Results

After adjusting for confounding factors, retrospective analysis revealed significant association between new-onset diabetes mellitus (NODM) and PC risk, with insulin treatment also linked to increased PC development. The standard inverse-variance weighted (IVW) method indicated that genetic susceptibility to T2DM was associated with an increased risk of developing PC (OR = 1.11; 95% CI = 1.034–1.193). Furthermore, MR showed T2DM, insulin treatment, FGF-4, and sulfhydryl oxidase 2 may be independently associated with the prevalence of PC. Specially, CMA demonstrated that insulin treatment, FGF4, and sulfhydryl oxidase 2 mediate the pathway from T2DM to PC, contributing 56.8%, 55.8%, and 5.9% of the total effect, respectively.

Conclusion

This study supports the association between T2DM, specifically NODM, and increased PC risk, with insulin therapy, FGF4, and sulfhydryl oxidase 2 mediating this pathway. Further research is required to elucidate the mechanisms underlying these mediating effects.

Clinical trial number

not applicable.

Peer Review reports

Background

Pancreatic cancer (PC) is a malignant tumor associated with low survival, with a median survival period of only 5–8 months and an outlier 5-year survival rate lower than 10% [1], causing a substantial global health burden [2]. Risk factors of PC primarily include personal attributes, lifestyle habits, metabolic disorders, and pancreatic disease status [3]. Notably, among these, type 2 diabetes mellitus (T2DM), a prevalent chronic noncommunicable disease, has shown complex bidirectional associations with PC [4].

Based on the time since onset, T2DM can be categorized as new-onset diabetes mellitus (NODM) or long-standing diabetes mellitus (LSDM) [5]. NODM, defined as a diagnosis within 24–36 months before PC surgery, is considered an early metabolic marker for PC [6], with numerous studies identifying a strong risk association between NODM and PC. In contrast, the relationship between LSDM and PC remains less consistent and varies both clinically and genetically [7, 8]. Glucose and lipid metabolism, especially glucose metabolism, and pancreatic disease status, could partially explain the connection between T2DM and PC [9]. Hyperinsulinemia and insulin resistance play key roles in PC progression, underscoring the influence of insulin in PC development [10, 11]. However, there is currently no direct research on the impact of T2DM therapies aimed at managing high blood glucose on PC risk. Although few studies suggest a higher risk of PC with insulin use, these findings are often confounded by bias [12]. Thus, T2DM therapy may be a pivotal mediator between T2DM and PC, which deserves further investigation.

Typical therapeutic strategies for T2DM include dietary management, oral medication, and insulin treatment. Recent research suggests that fibroblast growth factors (FGFs) may be involved in modulating various metabolic functions. Sun et al. reported that single intracerebroventricular administration of recombinant FGF4 can induce sustained T2DM remission in leptin receptor-deficient (db/db) mice [13]. Moreover, unlike FGF1 (a pan-FGFR ligand), FGF4 has no apparent effect on food intake. The potent anti-hyperglycemic and anti-inflammatory properties of FGF4 testify to its promising potential for use in the treatment of T2DM and related metabolic disorders [14, 15]. Some molecules involved in the FGF4 pathway, such as CSF-1, alpha-crystallin in B2, and gamma-crystallin in D, are known to be involved in cell proliferation, signaling, and differentiation, which may lead to the development of cancer [16,17,18,19]. However, FGF4 pathway has yet to be investigated for its mediating role between T2DM and PC.

Hence, in this article, we aimed to analyze the risk trends between T2DM and PC and assess their causal relationship through Mendelian randomization (MR) analysis. Additionally, we explored the mediating variables in the causal pathway to better understand the effects of T2DM on PC and the underlying mechanisms.

Methods

Statistical analysis of a single-center retrospective study

In this study, we included 581 adult patients (aged ≥ 18 years) who visited our center between January 2013 and December 2023 and were histopathologically confirmed with pancreatic cancer (PC) following surgery. Patients with other malignant tumors were excluded. Additionally, the Control Group comprised 582 participants, matched 1:1 with the patients by sex and age from the same hospital. All participants provided retrospective informed consent. The study was approved by the Ethics Committee of Hangzhou First People’s Hospital (Approval No. KY-20201114-0178-01), and all procedures followed ethical guidelines.

The study participants were divided into the pancreatic cancer group and the control group and were further classified into the T2DM group and the non-T2DM group based on the presence of type 2 diabetes mellitus (T2DM). For participants with T2DM, we collected data on the age at diagnosis (< 55 years, 55–65 years, ≥ 65 years), disease duration (≤ 1 year, 1–2 years, 2–5 years, 5–10 years, 10–20 years, ≥ 20 years), diabetes status [new-onset diabetes mellitus (NODM, disease duration ≤ 2 years) and long-standing diabetes mellitus (LSDM, disease duration > 2 years)], and treatment methods (diet control, oral medication, insulin therapy).

We documented each participant’s demographic information (e.g., age, sex), social determinants (e.g., marital status, education level), lifestyle factors (e.g., smoking and alcohol consumption), and health status (e.g., body mass index [BMI], history of coronary heart disease [CHD], and hypertension). Marital status was categorized as “married/remarried,” “unmarried/divorced,” and “widowed.” Education level was divided into three categories: primary school or below, junior high school, and high school or above. Smoking was defined as consuming more than one cigarette per day for more than six consecutive months, with smoking history classified as never smoked, former smoker, or current smoker. For current alcohol consumers, we inquired about specific details of alcohol use, including the type of beverage (spirits/beer/wine), drinking frequency (times per week), and quantity consumed per occasion. Excessive alcohol consumption was defined as a weekly intake of more than 210 g for men and more than 140 g for women. Moderate drinking was defined as consuming alcohol more than once per week for more than six consecutive months. Alcohol consumption was categorized as never consumed, former consumer, moderate drinker, or heavy drinker. We used a BMI threshold of > 24 to define an individual as overweight/obesity [20].

Baseline data were organized and analyzed using SPSS 26.0 statistical software. For normally distributed continuous data, results are expressed as xˉ±s, and comparisons between two groups were performed using independent samples t-test. Categorical data are presented as frequency (percentage), and comparisons between groups were conducted using chi-square test.

Multivariable logistic regression was performed to assess the association between T2DM and PC via odds ratios (ORs) and 95% confidence intervals (CIs). This study used disease status as the dependent variable (coded as 1 for presence and 0 for absence), and included the following as independent variables in the regression model: diabetes status, diabetes by age at diagnosis (years), diabetes by time since diagnosis (years), diabetes status by subtypes, diet, use of oral medication, and use of insulin. Three models were used for analysis. Model 1: Includes basic demographic information (age, sex) as covariates. Model 2: Adds clinical information (smoking, alcohol use, hypertension, obesity, coronary heart disease) as covariates to Model 1 Model 3: Further includes socioeconomic information (marital status, education level) as covariates to Model 2 All three models underwent Hosmer-Lemeshow testing, with P-values > 0.05, indicating good model fit. Additionally, collinearity diagnostics showed that all models met the criteria of tolerance > 0.01, VIF < 10, CI < 30, and eigenvalues > 0.01, suggesting that there were no collinearity issues among the covariates. Subsequently, mediating effects were assessed by incorporating an interaction term into the model and comparing it to a noninteractive model through likelihood ratio tests and subsequently validated through stratified analysis of these variables.

GWAS data sources for T2DM

T2DM-associated GWAS data are accessible from FinnGen (https://www.finngen.fi/en). As of January 2022, the project had amassed 309,154 samples, 16,962,023 variables, and 180,756 control participants. The T2DMGWAS included 215,654 participants and 16,380,440 single-nucleotide polymorphisms (SNPs) (Supplementary Table S1).

GWAS data sources for PC

The relevant GWASs for PC included 476,245 participants and 24,195,229 SNPs. The aggregated dataset is accessible for free via the IEU OpenGWAS (Supplementary Table S1).

GWAS data sources for the mediating factors

Mediating factors included lipid metabolism (Body Mass Index (BMI), total cholesterol, high density lipoprotein (HDL) cholesterol, triglycerides, low density lipoprotein (LDL) cholesterol), glucose metabolism (two-hour glucose challenge, fasting blood insulin, peak insulin response, insulin secretion rate, insulin levels, fasting glucose, insulin treatment), pancreatic function (pancreatic volume, acute pancreatitis (AP), chronic pancreatitis, liver, biliary or pancreas problem, other disorders of glucose regulation and pancreatic internal secretion), and the FGF4 pathway (FGF4, CSF-1, alpha-crystallin in B2, gamma-crystallin in D, PGC cross-disorder traits, sulfhydryl oxidase 2) [16,17,18,19]. The SNPs aggregated dataset is accessible for free via the IEU OpenGWAS (https://gwas.mrcieu.ac.uk/) (Supplementary Table S1). This study utilized deidentified data from ethically approved trials, exempting the need for individual ethical approval.

IV selection

The conventional Bonferroni-corrected threshold of p < 5 × 10− 8 was adopted as significant threshold, and the commonly used threshold of p < 1 × 10− 5 as a suggestive evidence level. When none SNPs reaches significant threshold, suggestive evidence level was applied to find more SNPs [21]. SNPs were evaluated utilizing rigorous criteria (10,000 kb, r2 ≤ 0.001). Outlier SNPs were identified using the MR-PRESSO test and subsequently deleted. We exclude SNPs with a minor allele frequency (MAF) less than 0.01, which indicates that enough individuals in the sample carry the allele, increasing the likelihood of detecting an association with the phenotype [22,23,24]. The F statistic was calculated as F = (R2 / (1 - R2)) * ((N - K − 1) / K), where R2 = 2 * MAF * (1 - MAF) * (β / SD)2 [25, 26], which is used to assess the strength of the SNPs. SNPs with F statistic < 10 were identified as weak IVs (Supplementary Table S2).

Mendelian randomization

Three fundamental assumptions must be upheld for MR: (1) correlation: the SNP demonstrates a significant association with the exposure factor; (2) independence: the SNP operates independently of any confounders; and (3) exclusion: the outcome is influenced by the SNP via only the exposure factor. After excluding SNPs with LD and those that were associated with other traits, a total of 306 SNPs were retained for genetic analysis. The inverse-variance weighted (IVW) method was applied to determine the causal effect of T2DM on PC, which combines the Wald ratio estimates of each SNP into 1 causal estimate for each exposure using the random-effects meta-analysis approach [27, 28]. In addition, sensitivity analyses were performed. Heterogeneity was assessed through Cochrane’s Q test, where a P-value < 0.05 would be considered an indication of heterogeneity. When significant heterogeneity was detected, a random-effects model was used; otherwise, a fixed-effects model was used. The MR Egger regression intercept analysis was conducted to examine horizontal pleiotropy, with a P-value < 0.05 considered as evidence of horizontal pleiotropy [25].

Bidirectional two-sample Mendelian randomization

We used the R software package (4.0.2), including two-sample MR and Mendelian Randomization. The causal impact of PC on T2DM was probed using identical procedures. SNPs with LD and those correlated with other traits were eliminated, yielding 37 significant SNPs for analysis (Supplementary Table S2).

Causal mediation analysis

If a causal relationship between T2DM and PC is established, two-step MR should subsequently be used to investigate the mediating effects of common risk factors for PC. Prevalent risk factors for PC were regarded as mediators, including glucose metabolism, lipid metabolism, pancreatic function and relevant phenotypes of the FGF4 pathway, with comprehensive data shown in the Supplementary Table S1. First, we estimated the causal effect of T2DM on these factors through two-sample MR to identify those factors significantly impacted by T2DM. Second, we identified causal risk factors for PC through two-sample MR. The overlapping traits of the first and second steps were identified as substantial mediating factors. The PhenoScanner V2 tool was used to detect SNPs pertaining to T2DM among the IVs. This ensures independence of the two sets of SNPs: T2DM (exposure) and risk factors (mediating factors). Using the product coefficient method, the indirect effect of T2DM on the mediating factors was measured by multiplying the influence of mediating factors on PC. The 95% CI for this indirect effect was estimated using the delta method.

Results

Baseline characteristics

Table 1 presents the baseline characteristics of the patients in the PC group and control group. Among the 1,163 participants, the patient group had a greater proportion of older individuals and was predominantly male (mean [standard deviation] age, 67.34 [55.32–79.36] years; male, 60.07%). Among the social determinants, a significantly higher percentage of individuals in the control group had junior high and higher education levels (41.92%), compared to the PC group (32.88%). The proportion of smokers/ex-smokers in the PC group was notably greater, while alcohol consumption was not significantly different between the two groups. Furthermore, there were fewer individuals with hypertension in the control group than in the PC group, yet there were more overweight individuals in the control group (182, 31.27%).

Table 1 Baseline data of the study population
Full size table

Effects of T2DM on PC

The associations between T2DM and PC risk are presented in Table 2. Across all models, patients with T2DM had a notably greater risk of developing PC than non-T2DM individuals. There was a downward trend in PC risk as T2DM duration increased. The risk of PC for NODM patients was greater than that for LSDM patients. Furthermore, the results suggested a trend toward increased cancer risk with age at T2DM diagnosis, peaking at > 65 years. Compared with T2DM patients managed through diet and oral medication, T2DM patients treated with insulin had a greater risk of PC.

Table 2 Association between T2DM-related variables and PC risk (581 patients and 582 control participants)
Full size table

Bidirectional two-sample MR analysis

Bidirectional two-sample MR was conducted to explore the relationship between T2DM and PC risk. A total of 306 SNPs were carefully selected to represent the genetic predisposition to T2DM (Supplementary Table S2). Using the standard IVW method, we found that genetic susceptibility to T2DM was associated with an increased risk of developing PC (OR = 1.11; 95% CI = 1.034–1.193; P = 0.0039) (Supplementary Table S3). The results of the MR analysis for T2DM and PC are illustrated in Fig. 1. In contrast, the analysis did not find a statistically significant causal relationship between PC and T2DM (Supplementary Table S3). Additionally, the MR-Egger intercepts indicated no evidence of directional pleiotropy (P > 0.05), and the analysis showed no significant heterogeneity (Supplementary Table S4).

Fig. 1
figure 1

Forest plot of the estimated results (ORs and 95% CIs) from the MR analysis. The point estimates are represented by a bullet along with the 95%CIs. (a) The directional association between T2DM and PC risk. (b) The directional association between PC and T2DM risk. IVW, inverse-variance weighted; MR ANALYSIS, MR analysis; PC, pancreatic cancer; T2DM, type 2 diabetes mellitus

Full size image

Causal mediation analysis

We performed CMA to assess the proportion of patients with T2DM influencing PC risk through modifiable risk factors. Figure 2 elucidates the causal impact of T2DM on prevalent risk factors, establishing a causal link with HDL, total cholesterol, decreased peak insulin response, diminished insulin levels, fasting glucose, T2DM, insulin treatment, other glucose regulations and pancreatic internal secretory disorders, FGF-4, and sulfhydryl oxidase2 (Supplementary Table S3, S5-6).

Fig. 2
figure 2

Separate MR associations between genetically predicted T2DM and each risk factor. IVW, inverse-variance weighted; PC, pancreatic cancer; T2DM, type 2 diabetes mellitus; AP, acute pancreatitis; LDL, low-density lipoprotein; HDL, high density lipoprotein; CSF, colony-stimulating factor; FGF, fibroblast growth factor

Full size image

Based on CMA, T2DM, insulin treatment, FGF-4, and sulfhydryl oxidase 2 may be independently associated with the prevalence of PC, which is unrelated to other risk factors (Fig. 3). We quantified the mediating effects of T2DM, insulin treatment, FGF-4, and sulfhydryl oxidase 2, as shown in Fig. 4. The mediating role of insulin treatment in the causal pathway from T2DM to PC was 0.061 (95% CI = 0.043–0.076), accounting for 56.8% of the total effect. The mediating role of FGF4 was 0.058 (95% CI = 0.023–0.095), contributing to 55.80% of the total effect. Sulfhydryl oxidase 2 mediated a causal pathway from T2DM to PC with an effect size of 0.006 (95% CI = 0.002–0.009), accounting for 5.90% of the total effect.

Fig. 3
figure 3

MR associations of each genetically predicted risk factor with PC. IVW, inverse-variance weighted; PC, pancreatic cancer; T2DM, type 2 diabetes mellitus; AP, acute pancreatitis; LDL, low-density lipoprotein; HDL, high-density lipoprotein; CSF, colony-stimulating factor; FGF, fibroblast growth factor

Full size image
Fig. 4
figure 4

MR-estimated effects of mediators. (a) MR-estimated effects of T2DM on each mediator, presented as β/ORs with 95% CIs. (b) MR-estimated effects of each mediator on PC, presented as β/ORs with 95% CIs. (c) MR-estimated effects of indirect effects of each mediator by the product of coefficients method with delta method-estimated 95%CIs. MR-estimated proportions (%) are presented with 95% CIs. T2DM, type 2 diabetes mellitus; PC, pancreatic cancer

Full size image

Discussion

In this retrospective study and bidirectional two-sample MR, we identified associations between T2DM subtypes and PC and revealed cancer risk among individuals with T2DM with diverse characteristics. This study is the first analysis identifying three mediating factors, including insulin treatment, FGF4 and sulfhydryl oxidase 2 between T2DM and PC via two-step MR.

The results from a bidirectional two-sample MR of individuals of European descent indicate that an increased OR for T2DM enhances susceptibility to PC, as based on the fixed-effects model. Inverse MR findings suggest the lack of a causal relationship between PC and T2DM, and similar outcomes were also observed in other research [29].The studies lacked evidence of a causal impact of T2DM on PC risk [30]. The discrepancy in results could be due to our use of recent T2DM and PC GWASs. Simultaneously, we employed multivariable regression to correct for confounding factors, and the retrospective study findings aligned with previous findings. A meta-analysis encompassing 88 retrospective studies revealed that the risk of PC increases in the initial years after the diagnosis of T2DM [31]. Consistently, the risk posed by T2DM persisting for less than two years is greater for NODM than for LSDM, suggesting the “diabetesizing” role of PC. As time progresses, the risk decreases and maintains an enduring association with PC. There is a temporal dependence between T2DM and PC. Our findings support the implementation of regular cancer screening in elderly patients who develop hyperglycemia within 2 years, providing guidance for glycemic control in elderly LSDM patients.

Beyond the identification of causality, potential mediators are also an area of research focus. The CMA suggested that insulin’s method of controlling blood glucose mediates the causal effect of T2DM on PC. The type of treatment for type 2 diabetes mellitus (T2DM) may depend on the severity of the condition. Severe T2DM is associated with an increased risk of pancreatic cancer, and this relationship may be influenced by the treatments administered for managing severe T2DM. It is important to consider how the severity of T2DM and its treatment could potentially mediate this risk. Previous research has indicated that physiological insulin levels facilitate pancreatic cell secretion of digestive enzymes for fat breakdown, yet high insulin levels can induce pancreatic inflammation and precancerous cell growth [32]. In the treatment of pancreatic cancer, researchers are committed to blocking the insulin /IGF signaling pathway in order to achieve cancer targeted therapy [33]. Therefore, from a public health perspective, patients with T2DM should utilize diet, exercise, and oral medication for optimal control. Careful consideration is required when initiating insulin therapy for T2DM to maintain insulin levels within the normal range.

FGF4 is a member of the fibroblast growth factor family, which plays a crucial role in various physiological and pathological processes. Previous studies have confirmed that FGF4 is important in regulating glucose homeostasis. For example, Sun et al. found that centrally administered FGF4 induced sustained remission in a T2DM mouse model, suggesting that FGF4 may regulate blood glucose levels by affecting glucose-sensing neurons in the brain [15]. Additionally, the role of FGF4 in promoting epithelial-mesenchymal transition (EMT), as well as the proliferation, migration/invasion, and colony formation of pancreatic cancer cells, highlighting the potential role of FGF4 in the progression of pancreatic cancer [34]. Qi et al. demonstrated that FGF4 increases intracellular calcium ion concentration by upregulating the expression of the calcium signaling-related protein Orai1, thereby promoting EMT [34]. Ying et al. found that FGF4 exerts a powerful antihyperglycemic effect by targeting skeletal muscle, where FGFR1c is highly expressed [14]. These findings suggest that FGF4 plays a critical role in both T2DM and PC. The dual role of FGF4 in T2DM and PC suggests that it may act as a bridge between the two diseases. On one hand, FGF4 may influence the progression of T2DM by regulating glucose homeostasis and insulin sensitivity; on the other hand, it may promote pancreatic cancer development by facilitating EMT and cellular proliferation. These findings help us better understand the biological link between T2DM and PC and provide clues for further exploration of potential therapeutic targets.

Based on the aforementioned studies, it can be speculated that FGF4 may play a role in the connection between T2DM and PC through the following mechanisms: In T2DM, FGF4 may influence disease progression by regulating glucose homeostasis and insulin sensitivity; in PC, FGF4 may promote tumor development by facilitating EMT and cell proliferation; and FGF4 may affect the behavior of pancreatic cancer cells by modulating intracellular calcium ion concentrations, which offers potential avenues for the development of new therapeutic strategies.

While FGF4 provides valuable insights into the T2DM-PC connection, another important molecule, Quiescin sulfhydryl oxidase (QSOX), also warrants attention for its role in cancer progression. QSOX is a unique, multi-domain disulphide catalyst that is localized primarily to the Golgi apparatus and secreted fluids and has attracted attention owing to its overproduction in tumors [35]. Sulfhydryl oxidase1 activity can be associated with metastasis or progression in several cancers, such as breast cancer and pancreatic ductal adenocarcinoma [36]. Mechanistically, QSOX1 post-translationally activates matrix metalloproteinases (MMPs) [37]. Although whether sulfhydryl oxidase 2 is involved in tumor growth has not been revealed yet, the regulatory role of sulfhydryl oxidase2 in PC development should not be ignored via its similar sequence and structure to sulfhydryl oxidase 1. But no studies explored its role in T2DM. Our study emphasized for the first time that sulfhydryl oxidase2 may plays a mediatory role in association between T2DM and PC. This finding suggests that sulfhydryl oxidase2 could also be a potential therapeutic target for PC invasion, providing a valuable direction for treatment.

Moreover, our retrospective study and MR results refute an association between excess body weight and PC risk, as participant-reported body weight measures might introduce common biases in retrospective studies, such as survival bias or overweight misclassification. Cancer is often characterized by symptoms of weight loss and reduced body mass prior to or early in diagnosis, a factor possibly accounting for previous MR findings [38].

There are still some limitations of our research. First, retrospective studies risk recall bias, selection bias, and information bias in data collection, thereby reducing the reliability of the study outcomes. Second, due to limited data on MR in individuals of east Asian descent, MR research was conducted solely among European populations, necessitating caution when applying its results to other groups. Larger-scale GWASs in individuals of East Asian descent are needed to reassess outcomes in the future. Due to the lack of relevant GWAS data for T2DM subtypes, the associations of NODM and LSDM with PC can only be deduced from retrospective studies. In addition, two-step MR analysis may not adequately control for potential interplay between exposure and mediators, causing bias. In summary, identifying FGF4 and sulfhydryl oxidase 2 as potential mediators in the relationship between T2DM and PC provides us with a new perspective, helping us to better understand the association between these two diseases and offering potential targets for future therapeutic strategies. These findings advance our understanding of disease processes and offer significant directions for further research.

Conclusions

Our study identified the high risk of developing PC in T2DM patients, suggesting that insulin therapy may play a mediating role in the causal pathway between T2DM and PC, which potentially offers a reliable reference for glycemic management protocols in T2DM patients. The results are limited by multiple factors and may need to be confirmed in later studies. Moreover, the findings regarding the mediatory role of FGF4 in the above association demonstrated its potential as a predictive biomarker of PC.

Data availability

All data generated or analysed during this study are included in this published article.

Abbreviations

AP:

acute pancreatitis

BMI:

body mass index

CHD:

coronary heart disease

CSF-1:

colony-stimulating factor

ECM:

extracellular matrix proteins

FGF4:

fibroblast growth factor 4

FGFR:

fibroblast growth factor receptor

GSN:

glucose-sensitive neuron

GWAS:

genome-wide association study

HDL:

high-density lipoprotein

IV:

instrumental variable

IVW:

inverse-variance weighted

LD:

linkage disequilibrium

LDL:

low-density lipoprotein

LSDM:

long-standing diabetes mellitus

MMP:

matrix metalloproteinase

MR:

Mendelian randomization

CMA:

causal mediation analysis

NODM:

new-onset diabetes mellitus

OR:

odds ratio

PC:

pancreatic cancer

SNP:

single-nucleotide polymorphism

T2DM:

type 2 diabetes mellitus

References

  1. Klein AP. Pancreatic cancer epidemiology: Understanding the role of lifestyle and inherited risk factors. Nat Rev Gastroenterol Hepatol. 2021;18(7):493–502.

    PubMed  PubMed Central  Google Scholar 

  2. Zhao C, Gao F, Li Q, Liu Q, Lin X. The distributional characteristic and growing trend of pancreatic cancer in China. Pancreas. 2019;48(3):309–14.

    PubMed  Google Scholar 

  3. Stoffel EM, Brand RE, Goggins M. Pancreatic cancer: changing epidemiology and new approaches to risk assessment, early detection, and prevention. Gastroenterology. 2023;164(5):752–65.

    PubMed  Google Scholar 

  4. Ayan R, Jayaprakash S, Sadishkumar K, Dukhabandhu N, Pazhanivel M, Raja K. Diabetes and pancreatic cancer: exploring the two-way traffic. World J Gastroenterol. 2021;27(30):4939–62.

    Google Scholar 

  5. Ahlqvist E, Storm P, Krjmki A, Martinell M, Dorkhan M, Carlsson A, Vikman P, Prasad RB, Aly DM, Almgren P. Novel subgroups of adult-onset diabetes and their association with outcomes: a data-driven cluster analysis of six variables. Lancet Diabetes Endocrinol 2018:S2213858718300512.

  6. Hart PA, Bellin MD, Andersen DK, Bradley D, Cruz-Monserrate Z, Forsmark CE, Goodarzi MO, Habtezion A, Korc M, Kudva YC. Type 3c (pancreatogenic) diabetes mellitus secondary to chronic pancreatitis and pancreatic cancer. Lancet Gastroenterol Hepatol. 2016;1(3):226–37.

    PubMed  PubMed Central  Google Scholar 

  7. Dunne RF, Roeland EJ. The interplay among pancreatic cancer, cachexia, body composition, and diabetes. Hematol Oncol Clin North Am. 2022;36(5):897–910.

    PubMed  Google Scholar 

  8. Pandol SJ, Forsmark CE, Consortium for the Study of Chronic, Pancreatitis D, Pancreatic C. Consortium for the study of chronic pancreatitis, diabetes, and pancreatic cancer: achievements and future directions. Curr Opin Gastroenterol. 2021;37(5):486–90.

    PubMed  PubMed Central  Google Scholar 

  9. Singhi AD, Koay EJ, Chari ST, Maitra A. Early detection of pancreatic cancer: opportunities and challenges. Gastroenterology. 2019;156(7):2024–40.

    PubMed  Google Scholar 

  10. Li J, Cao G, Ma Q, Liu H, Li W, Han L. The bidirectional interation between pancreatic cancer and diabetes. World J Surg Oncol. 2012;10(1):171.

    CAS  PubMed  PubMed Central  Google Scholar 

  11. Suryadevara V, Roy A, Sahoo J, Kamalanathan S, Naik D, Mohan P, Kalayarasan R. Incretin based therapy and pancreatic cancer: realising the reality. World J Gastroenterol. 2022;28(25):2881–9.

    CAS  PubMed  PubMed Central  Google Scholar 

  12. Katarina P, Brigita S, Milica M, Ljiljana V, THE RELATIONSHIP BETWEEN DIABETES MELLITUS AND PANCREATIC CANCER - DIABETES MELLITUS AS A RED FLAG FOR PANCREATIC CANCER. Cancer epidemiology, biomarkers & prevention: a publication of the American association for cancer research. Cosponsored Am Soc Prev Oncol 2023, 32(3).

  13. Sundaram SM, Lenin RR, Janardhanan R. FGF4 alleviates hyperglycemia in diabetes and obesity conditions. Trends Endocrinol Metab. 2023;34(10):583–5.

    CAS  PubMed  Google Scholar 

  14. Ying L, Wang L, Guo K, Hou Y, Li N, Wang S, Liu X, Zhao Q, Zhou J, Zhao L, et al. Paracrine FGFs target skeletal muscle to exert potent anti-hyperglycemic effects. Nat Commun. 2021;12(1):7256.

    CAS  PubMed  PubMed Central  Google Scholar 

  15. Sun H, Lin W, Tang Y, Tu H, Chen T, Zhou J, Wang D, Xu Q, Niu J, Dong W, et al. Sustained remission of type 2 diabetes in rodents by centrally administered fibroblast growth factor 4. Cell Metabol. 2023;35(6):1022–e10371026.

    CAS  Google Scholar 

  16. Ghosh JG, Shenoy AK Jr., Clark JI. Interactions between important regulatory proteins and human AlphaB Crystallin. Biochemistry. 2007;46(21):6308–17.

    CAS  PubMed  Google Scholar 

  17. Fujimoto M, Izu H, Seki K, Fukuda K, Nishida T, Yamada S, Kato K, Yonemura S, Inouye S, Nakai A. HSF4 is required for normal cell growth and differentiation during mouse lens development. EMBO J. 2004;23(21):4297–306.

    CAS  PubMed  PubMed Central  Google Scholar 

  18. Kawase E, Hashimoto K, Pedersen RA. Autocrine and paracrine mechanisms regulating primordial germ cell proliferation. Mol Reprod Dev. 2004;68(1):5–16.

    CAS  PubMed  Google Scholar 

  19. Aglan HA, Kotob SE, Mahmoud NS, Kishta MS, Ahmed HH. Bone marrow stem cell-derived β-cells: new issue for diabetes cell therapy. Tissue Cell. 2024;86:102280.

    CAS  PubMed  Google Scholar 

  20. Pan XF, Wang L, Pan A. Epidemiology and determinants of obesity in China. Lancet Diabetes Endocrinol. 2021;9(6):373–92.

    PubMed  Google Scholar 

  21. Li Z, Wang W, Tian X, Duan H, Xu C, Zhang D. Bivariate genome-wide association study (GWAS) of body mass index and blood pressure phenotypes in Northern Chinese twins. PLoS ONE. 2021;16(2):e0246436.

    CAS  PubMed  PubMed Central  Google Scholar 

  22. Huang W, Zou Y, Zhang K, Yao S, Tang SH, Wu H, Wang PF, Xue HZ, Yang TL, Zhang K et al. Two-sample Mendelian randomization analysis reveals causal relationships between blood lipids and venous thromboembolism. Haematologica 2024.

  23. Li P, Wang H, Guo L, Gou X, Chen G, Lin D, Fan D, Guo X, Liu Z. Association between gut microbiota and preeclampsia-eclampsia: a two-sample Mendelian randomization study. BMC Med 2022.

  24. Cloé D, PierreEmmanuel S, Kumar SAA, Claudia S, Sandeep G, Océane M, Berta P, Patrick M, Milena RBD. Mendelian randomisation study of smoking, alcohol, and coffee drinking in relation to Parkinson’s disease. J Parkinson’s Disease. 2021;12(1):267–82.

    Google Scholar 

  25. Ke TM, Lophatananon A, Muir KR. Strengthening the evidence for a causal link between type 2 diabetes mellitus and pancreatic cancer: insights from Two-Sample and multivariable Mendelian randomization. Int J Mol Sci 2024, 25(9).

  26. Xie Q, Hu B. Effects of gut microbiota on prostatic cancer: a two-sample Mendelian randomization study. Front Microbiol. 2023;14:1250369.

    PubMed  PubMed Central  Google Scholar 

  27. Yao S, Zhang M, Dong SS, Wang JH, Zhang K, Guo J, Guo Y, Yang TL. Bidirectional two-sample Mendelian randomization analysis identifies causal associations between relative carbohydrate intake and depression. Nat Hum Behav. 2022;6(11):1569–76.

    PubMed  Google Scholar 

  28. Wang Y, Ye C, Kong L, Zheng J, Xu M, Xu Y, Li M, Zhao Z, Lu J, Chen Y, et al. Independent associations of education, intelligence, and cognition with hypertension and the mediating effects of cardiometabolic risk factors: A Mendelian randomization study. Hypertens (Dallas Tex: 1979). 2023;80(1):192–203.

    CAS  Google Scholar 

  29. Ye L, Manuel G, Justo L-B, Luca M, Ofure O, Daniele C, Federico C. Mendelian randomisation study of the effects of known and putative risk factors on pancreatic cancer. J Med Genet. 2020;57(12):jmedgenet–2019.

    Google Scholar 

  30. Robert CT, Mattias J, Valerie G, Haycock PC, Wade KH, Relton CL, Martin RM, George DS, Paul B. The role of obesity, type 2 diabetes, and metabolic factors in pancreatic cancer: A Mendelian randomization study. J Natl Cancer Inst 2017(9):495.

  31. Pikli B, Stephen VH, Christopher C, Mehrdad N. Association of diabetes mellitus and pancreatic adenocarcinoma: a meta-analysis of 88 studies. Ann Surg Oncol. 2014;21(7):2453–62.

    Google Scholar 

  32. Y ZAM, Han XY, H LJS, K WWC HCK, Nan JRTJCYJC, Justin C, Shilpa C. Hyperinsulinemia acts via acinar insulin receptors to initiate pancreatic cancer by increasing digestive enzyme production and inflammation. Cell Metabol. 2023;35(12):2119–e21352115.

    Google Scholar 

  33. Ceren MA, Erdinc BH, Shenghan W, Helmut F, O CG, Ekin DI. Insulin/IGF-driven cancer cell-stroma crosstalk as a novel therapeutic target in pancreatic cancer. Mol Cancer 2018, 17(1).

  34. Qi L, Song W, Li L, Cao L, Yu Y, Song C, Wang Y, Zhang F, Li Y, Zhang B, et al. FGF4 induces epithelial-mesenchymal transition by inducing store-operated calcium entry in lung adenocarcinoma. Oncotarget. 2016;7(45):74015–30.

    PubMed  PubMed Central  Google Scholar 

  35. Alon A, Grossman I, Gat Y, Kodali VK, DiMaio F, Mehlman T, Haran G, Baker D, Thorpe C, Fass D. The dynamic disulphide relay of quiescin sulphydryl oxidase. Nature. 2012;488(7411):414–8.

    CAS  PubMed  PubMed Central  Google Scholar 

  36. Li Y, Liu M, Zhang Z, Deng L, Zhai Z, Liu H, Wang Y, Zhang C, Xiong J, Shi C. QSOX2 is an E2F1 target gene and a novel serum biomarker for monitoring tumor growth and predicting survival in advanced NSCLC. Front Cell Dev Biology. 2021;9:688798.

    Google Scholar 

  37. Lake DF, Faigel DO. The emerging role of QSOX1 in cancer. Antioxid Redox Signal. 2014;21(3):485–96.

    CAS  PubMed  PubMed Central  Google Scholar 

  38. Molina-Montes E, Coscia C, Gómez-Rubio P, Fernández A, Ennis D. Deciphering the complex interplay between pancreatic cancer, diabetes mellitus subtypes and obesity/bmi through causal inference and mediation analyses. Gut 2020:gutjnl–2019.

Download references

Acknowledgements

None.

Funding

This study was supported in part by the Zhejiang Province Key R&D Plan Project (Grant Nos. 2023C03054 and 2024C03048); the Natural Science Foundation of Zhejiang Province (Grant No. LQ24H030008); the Zhejiang Medical and Health Science and Technology Plan (Grant Nos. 2022RC056, 2023RC229 and 2024KY185); the Zhejiang Chinese Medicine Science and Technology Plan (Grant Nos. GZY-ZJ-KJ-24093 and 2024ZF112); the Hangzhou Medical and Health Science and Technology Plan (Grant No.A20230021); and the Construction Fund of Medical Key Disciplines of Hangzhou (OO20190001).

Author information

Authors and Affiliations

  1. The Fourth School of Clinical Medicine, Zhejiang Chinese Medical University, Hangzhou, China

    Yuxin Wang & Xiaofeng Zhang

  2. Department of Gastroenterology, Affiliated Hangzhou First People’s Hospital, Westlake University School of Medicine, Hangzhou, China

    Lu Xie, Ye Gu, Hangbin Jin, Jianfeng Yang, Qiang Liu & Xiaofeng Zhang

  3. Key Laboratory of Integrated Traditional Chinese and Western Medicine for Biliary and Pancreatic Diseases of Zhejiang Province, Hangzhou, 310006, China

    Jianfeng Yang, Qiang Liu & Xiaofeng Zhang

  4. Key Laboratory of Clinical Cancer Pharmacology and Toxicology Research of Zhejiang Province, Hangzhou, 310058, China

    Jianfeng Yang & Xiaofeng Zhang

  5. Hangzhou Hospital & Institute of Digestive Diseases, Hangzhou, Hangzhou, 310006, China

    Jianfeng Yang & Xiaofeng Zhang

Authors
  1. Yuxin Wang
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Lu Xie
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Ye Gu
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Hangbin Jin
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Jianfeng Yang
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Qiang Liu
    View author publications

    You can also search for this author inPubMed Google Scholar

  7. Xiaofeng Zhang
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

WYX is first authors. LQ, and ZXF designed the study. XL, GY, JHB, and YJF collected the data. WYX analyzed the data. WYX drafted the manuscript. XL and GY contributed to the interpretation of the results and critical revision of the manuscript for important intellectual content and approved the final version of the manuscript. All authors have read and approved the final manuscript.

Corresponding authors

Correspondence to Qiang Liu or Xiaofeng Zhang.

Ethics declarations

Ethics approval and consent to participate

All procedures were performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. The study was approved by the Ethics Committee of Hangzhou First People’s Hospital (Approval No. KY-20201114-0178-01). All participants provided written informed consent. The study was carried out in accordance with the applicable guidelines and regulations.

Consent for publication

Not Applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Supplementary Material 2

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Wang, Y., Xie, L., Gu, Y. et al. Complex interplay between type 2 diabetes mellitus and pancreatic cancer: insights from observational and mendelian randomization analyses. BMC Cancer 25, 556 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12885-025-13976-6

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12885-025-13976-6

Keywords

BMC Cancer

ISSN: 1471-2407

Contact us