Dreadful Gastric Cancer And Mortality Risk After Diabetes Diagnosis
Overview
Evidence indicates a potential link between diabetes and the risk of gastric cancer, but the conclusions are not definitive, especially with limited research focusing on Asian populations. This study aimed to evaluate the impact of diabetes and its duration on the overall development of gastric cancer, considering different anatomical and histological subtypes.
The analysis pooled data from 12 prospective studies within the Asia Cohort Consortium. The study included 558,981 participants with a median age of 52. Over a median follow-up period of 14.9 years, 8,556 incident primary gastric cancers and 8,058 gastric cancer deaths were recorded. Cox proportional hazard regression models estimated study-specific hazard ratios (HRs) and 95% confidence intervals (CIs), which were then pooled using random-effects meta-analyses.
Diabetes was found to be associated with an increased overall incidence of gastric cancer (HR 1.15, 95% CI 1.06–1.25). This association did not significantly vary by sex (women vs men: HR 1.31, 95% CI 1.07–1.60 vs 1.12, 1.01–1.23), anatomical subsite (noncardia vs cardia: 1.14, 1.02–1.28 vs 1.17, 0.77–1.78), or histological subtype (intestinal vs diffuse: 1.22, 1.02–1.46 vs 1.00, 0.62–1.61). The risk of gastric cancer was significantly higher during the first decade following a diabetes diagnosis (HR 4.70, 95% CI 3.77–5.86) and decreased over time (nonlinear p < .01). Additionally, positive associations were found between diabetes and gastric cancer mortality (HR 1.15, 95% CI 1.03–1.28), though these associations weakened after a 2-year time lag.
Diabetes is associated with a higher incidence of gastric cancer, regardless of sex, anatomical subsite, or cancer subtype. The risk is particularly elevated during the first decade after a diabetes diagnosis.
Introduction
Asian countries, especially those in East Asia, have the highest rates of gastric cancer, despite a decline in recent years. Conversely, diabetes prevalence is rising globally, particularly in Asia, home to over 60% of those with the disease. Diabetes is associated with increased risks of various cancers, including liver, pancreatic, endometrial, colorectal, breast, and bladder cancers, likely due to factors such as hyperinsulinemia, hyperglycemia, and chronic inflammation. However, the connection between diabetes and gastric cancer remains unclear.
Numerous studies suggest a positive correlation between diabetes and gastric cancer risk, while others have not found such a link. Some research indicates this association in both men and women, whereas others report it only in women. Few studies have examined the relationship by anatomical subsites of gastric cancer, yielding inconclusive results. In Asia, research from Korea, Japan, and Hong Kong has shown conflicting findings regarding the diabetes-gastric cancer link. These studies often suffer from limitations, such as small sample sizes, inadequate control of confounding factors, failure to stratify by sex, and insufficient analysis of cancer subtypes. Additionally, while some studies consider diabetes duration, none have explored its dose-response effect.
Diabetes in Asians exhibits unique epidemiological traits compared to Western populations, possibly due to early beta-cell dysfunction and impaired insulin action associated with low lean mass. Understanding the diabetes-gastric cancer relationship in Asians could help identify differential risks, leading to improved prevention strategies and surveillance for diabetic patients.
This study within the Asia Cohort Consortium aims to clarify the association between diabetes and gastric cancer in the Asian population, considering sex, anatomical subsites, histological subtypes, and diabetes duration.
Method
The Asia Cohort Consortium is an ongoing international collaborative study designed to explore the relationship between genetic and environmental factors and the onset of various diseases and conditions. This large-scale study includes over 1 million healthy participants from 40 cohorts across 10 Asian countries. For this specific research, 12 cohorts from China (Shanghai Men’s Health Study [SMHS], Shanghai Women’s Health Study [SWHS]), Japan (Three-Prefecture Cohort Study Aichi [Aichi3P], Japan Collaborative Cohort Study [JACC], Japan Public Health Center-Based Prospective Study [JPHC I, II], Life Span Study [LSS], Takayama Study), Korea (Korea Multicenter Cancer Cohort Study [KMCC], Korea National Cancer Center Cohort Study [KNCC], Seoul Male Cancer Cohort [SeoulM]), and India (Mumbai Cohort Study [MCS]) agreed to participate. Detailed characteristics of each cohort are provided in Table S1. Data from these cohorts were harmonized and anonymized at the Consortium’s coordinating center.
The study was ethically approved by the relevant committees overseeing each cohort and the institutional review boards at the Asia Cohort Consortium coordinating centers (Fred Hutchinson Cancer Research Center, Seattle, WA, USA; National Cancer Center, Tokyo, Japan). All participating cohorts required data on gastric cancer incidence and/or mortality and baseline self-reported diabetes diagnosis. Participants with missing information on age, sex, follow-up time, gastric cancer outcome, diabetes status, or previous cancer diagnoses at baseline were excluded. The final analyses included 558,981 participants from 11 cohorts for incidence analysis and 580,663 participants from 12 cohorts for mortality analysis.
For exposure assessment, self-reported diabetes diagnosis data were collected at recruitment, with previous studies validating these self-reports against glycosuria results. Age at diabetes diagnosis was available in five cohorts, and diabetes duration was calculated using three methods: sum of duration at recruitment and follow-up time, years between diagnosis age and enrollment age, and for newly diagnosed diabetes (<1 year at enrollment). Information on diabetes type, blood glucose levels, or medication use was unavailable.
Outcome ascertainment involved linking cancer cases to death certificates, cancer registries, vital statistics registries, or through active follow-up surveys, depending on each cohort’s protocol. Gastric cancers were classified using the International Classification of Diseases (ICD) codes, and anatomical subsite and histology information was available in certain cohorts for incidence analysis.
Covariates such as demographics, lifestyle factors, and anthropometry were collected via questionnaires at enrollment, with BMI calculated as weight in kilograms divided by height in meters squared. Smoking and alcohol consumption were categorized as never or ever, except in the Takayama cohort where current alcohol consumption (yes/no) was used. Missing values were imputed using the median for continuous variables and the mode for categorical variables, as the proportion of missing values was less than 5% in each cohort.
Statistical Analysis
Statistical analysis involved summarizing baseline characteristics using means, medians, ranges, and standard deviations for continuous variables and percentages for categorical variables. Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated using Cox proportional hazards regression for each cohort and pooled using random-effects meta-analysis. Time at entry was defined as the age at enrollment, and time at exit was the age at first primary gastric cancer diagnosis or death, any-cause death, loss to follow-up, or end of follow-up, whichever occurred first. Heterogeneity among cohorts was assessed using the I2 index. Models were adjusted for smoking status, alcohol consumption, BMI, and education level as available in each cohort. Stratified analyses were performed by sex, country, enrollment year, age at enrollment, age at diagnosis/death, and birth year.
Secondary analyses included examining gastric cancer anatomical subsites and histological subtypes, and analyzing diabetes duration by quartiles. Sensitivity analyses excluded gastric cancer cases within the first two years of follow-up and restricted participants to those over 30 years old to minimize the inclusion of type 1 diabetes cases. Competing risks models were developed to investigate the impact of diabetes and diabetes duration on gastric cancer risk.
To provide an overview of past research on the association between diabetes and gastric cancer, a systematic search using PubMed was conducted, covering articles published between 2000 and 2022. Studies assessing the risk of gastric cancer incidence due to diabetes and reporting rate ratios were included for meta-analysis and subgroup analysis. Random-effect models, I2 statistics, Egger’s regression test, and funnel plots were used to assess estimated effects, statistical heterogeneity, and publication bias. Statistical significance was set at p < .05, and analyses were performed using SAS 9.4 and Stata 16.0.
Result
During a median follow-up of 14.7 years, 8,556 cases of first primary gastric cancer were identified (5,676 men and 2,880 women) among participants with a median age of 52 years and a mean BMI of 23.1 kg/m². Of these participants, 36% had a history of smoking, 43% had consumed alcohol, and 4.8% reported having diabetes.
Individuals with a history of diabetes showed a 15% increased risk of developing gastric cancer (HR 1.15, 95% CI 1.06–1.25), with variations observed across different cohorts. For example, the KMCC cohort indicated a significant association (HR 1.50, 95% CI 1.04–2.16), while the JACC cohort showed a nonsignificant inverse association (HR 0.85, 95% CI 0.67–1.09). The risk of gastric cancer was higher in men by 12% (HR 1.12, 95% CI 1.01–1.23) and even higher in women by 31% (HR 1.31, 95% CI 1.07–1.60).
Country-specific analysis revealed that Korea had the highest association for overall gastric cancer (HR 1.32, 95% CI 1.01–1.71). The risk was notably higher in individuals who enrolled after 1990, those born between 1930 and 1949, those enrolled in their 50s, and those diagnosed in their 60s. Subsite analysis showed diabetes was significantly associated with noncardia gastric cancer (HR 1.14, 95% CI 1.02–1.28) but not with cardia gastric cancer (HR 1.17, 95% CI 0.77–1.78). A positive association was observed for the intestinal subtype (HR 1.22, 95% CI 1.02–1.46), but not for the diffuse subtype (HR 1.00, 95% CI 0.62–1.61).
The risk of gastric cancer increased significantly within the first decade of diabetes (HR 4.70, 95% CI 3.77–5.86) but decreased thereafter, forming an inverted V or lambda-shaped curve. When diabetes duration was assessed at baseline, the risk increased by 40% in the second quartile (median = 2.7 years; HR 1.40, 95% CI 1.01–1.95) and by 39% in the third quartile (median = 5.6 years; HR 1.39, 95% CI 1.00–1.93).
Analysis of gastric cancer mortality included 580,663 participants and 5,084 deaths during the follow-up. Diabetes was associated with a 15% increased risk of death from gastric cancer (HR 1.15, 95% CI 1.03–1.28). Subgroup analyses revealed a significant association for cardia gastric cancer mortality (HR 2.18, 95% CI 1.14–4.14) but not for noncardia gastric cancer mortality (HR 1.08, 95% CI 0.95–1.24).
In sensitivity analyses, the association between diabetes and gastric cancer incidence and mortality remained consistent, even after excluding cases within the first two years of follow-up. Competing risk analyses showed a slight decrease in the association for gastric cancer incidence, which remained significant, while the estimate for mortality became nonsignificant. The inverted lambda-shaped relationship between diabetes duration and gastric cancer risk persisted.
A meta-analysis of 26 studies confirmed that diabetes was associated with a higher risk of overall gastric cancer (RR 1.24, 95% CI 1.15–1.34) with significant heterogeneity across the studies. Subgroup analyses showed no significant differential associations by sex, region, study design, or other factors.
Conclusion
In a pooled analysis of Asian populations, our study identified a significant association between diabetes and an increased risk of both gastric cancer incidence and mortality. This comprehensive assessment, the largest of its kind, explored the relationship between diabetes duration and gastric cancer within the Asian demographic.
Multiple mechanisms may explain this link. Elevated insulin and hyperglycemia can directly promote gastric cancer by activating tumorigenic pathways such as PI3K/AKT/mTOR and insulin-like growth factor-1 (IGF-1), which is closely associated with insulin resistance. IGF-1 fosters gastric cancer cell proliferation and invasion through the Ras/RAF/MEK/ERK and JAK1/STAT3 signaling pathways. Hyperglycemia enhances gastric carcinoma proliferation and migration via the Pin1/BRD4 transcriptional pathway, contributing to the hyperglycemia-induced inflammatory process. Additionally, hyperglycemia sustains H. pylori infection, a known risk factor for gastric cancer, allowing the bacteria to thrive in diabetic patients. H. pylori’s cytotoxin-associated gene A antigen, an oncoprotein, shows increased adhesion, growth, and viability under high glucose conditions. Certain glucose-lowering therapies, such as sulfonylureas and insulin therapy, may elevate gastric cancer risk by interacting with insulin and IGF-1R signaling. Conversely, metformin, which mitigates insulin resistance, may reduce this risk by activating AMPK and inhibiting IGF-1R.
Our findings support previous studies indicating a positive association between diabetes and gastric cancer risk. Our meta-analysis, which included subgroup analyses by sex, anatomical subsite, study type, and follow-up duration, confirmed this association. Unlike earlier meta-analyses, we focused on studies published post-2000 for temporal relevance and conducted detailed subgroup analyses of gastric cancer subsites (cardia and non-cardia), addressing a gap in prior research. The follow-up period was also extended, allowing for a more nuanced understanding of diabetes’ impact on gastric cancer over time. Although heterogeneity among studies was noted, it was relatively low among the Asian cohorts in our analysis.
We observed a stronger association between diabetes and gastric cancer in women than in men, though this difference was not statistically significant. Our meta-analysis did not find significant differences between cardia and non-cardia gastric cancer regarding their association with diabetes. However, cardia gastric cancer showed a slightly higher positive association. This might be influenced by the relatively small number of cardia gastric cancer cases. Literature suggests that H. pylori infection, more commonly linked to non-cardia gastric cancers, may interact with diabetes, while obesity, often associated with cardia cancers, may also interact with diabetes. Moreover, the positive association was noted only with intestinal gastric cancer, likely due to the stronger link with H. pylori infection.
The association between diabetes and gastric cancer mortality was reevaluated within the Asian Cohort Consortium. Although a positive association was found, it disappeared in the 2-year lag time and competing risk analyses, indicating diabetes may not be a primary risk factor for gastric cancer mortality. The risk of gastric cancer was highest within the first 10 years of diabetes diagnosis, possibly due to hyperglycemia-hyperinsulinemia. After 11 years, the association weakened, similar to findings from a US study, possibly due to reduced endogenous insulin production over time.
Strengths of our study include a large sample size, long-term follow-up, and detailed individual-level data, allowing for comprehensive analyses and adjustments for potential confounders. However, limitations include the lack of H. pylori infection data, potential misclassification of diabetes due to self-reporting, and the inability to control for all gastric cancer-related factors. Despite these, our findings suggest diabetes significantly increases gastric cancer risk, particularly within the first decade of diagnosis. Given the high prevalence of diabetes in Asia, our results highlight the importance of maintaining metabolic health and early-stage diabetes surveillance to prevent gastric cancer.
Oncology Related Tools
Other
Latest Research
