In the present study, we used a two-sample Mendelian randomization approach and utilized data from the OPEN GWAS database to conduct a comprehensive investigation of the potential causal association between insulin use and digestive system tumors. The findings suggest that insulin use is significantly associated with an increased risk of CRC, but the current data cannot prove a causal link between insulin use and other digestive system cancers.
With the rapid development of human economy and science and technology, people's lifestyles are constantly changing. Diets high in sugar have led to an increasing incidence of diabetes [36]. The high incidence and high cost of treatment make diabetes an increasingly serious global public health problem, posing a great threat to people's health [37]. According to statistics, the number of patients with diabetes in the world was about 425 million in 2017, and the prevalence reached 9.3% in 2019. It is expected that by 2045, 629 million people will be troubled by health problems caused by diabetes, and the number of patients in 2045 will increase by 50% compared with 2017 [17]. Diabetes mellitus is one of the most common endocrine diseases [18], which is mainly divided into type 1 diabetes mellitus (T1D) and type 2 diabetes mellitus (T2D) according to different causes, among which T2D accounts for up to 90% of diabetic patients. The main pathogenesis of T1D is due to the immune system attack on pancreatic β-cell, resulting in insulin levels that cannot meet the physiological needs of the body, mainly in adolescents. On the contrary, T2D is common in middle-aged and elderly patients and is often accompanied by impairment of pancreatic β-cell function. Therefore, almost all patients with diabetes eventually require exogenous insulin to fill the physiological gap. According to the blood glucose control guidelines of the American Diabetes Association and the results of a number of cohort studies, the early introduction of insulin therapy in diabetic patients can reduce the occurrence of microangiopathy [19–21]. Therefore, the use of exogenous insulin is becoming more and more common in the treatment of patients with diabetes, and the global use of insulin is showing an upward trend.
Whether the use of insulin increases the risk of cancer is a difficult question that has long puzzled people and has attracted extensive attention from medical researchers. Since patients with diabetes are almost invariably treated with insulin, understanding the causal relationship between insulin and cancer is particularly important. At present, there are two conjectures about the mechanism of insulin promoting tumor growth, the first one is that insulin is related to the binding of insulin receptor or IGF receptor [28, 31]. Another major mechanism is that insulin promotes the binding of IGF-1 to the IGF receptor by inhibiting IGF-binding proteins. An interactive network is formed among IGF family members and among other growth factor families through various signaling pathways, thus playing an important role in cell proliferation, differentiation, apoptosis and transformation [29]. Among them, IGF family members have been shown to be closely related to the progression and metastasis of cancer [27, 30]. However, there are no studies that can clarify the complex pathways and pathogenesis between insulin use and cancer, so people can only speculate on the relationship between the two through epidemiological and animal studies [8–13, 22, 23]. In fact, even though these cohort studies and animal experiments have shown that insulin may increase the risk of cancer [24], there is still a large deviation from the actual clinical practice. Too small sample size, different populations, inconsistent follow-up time, and changes in dietary habits can lead to bias in the final results.
Therefore, we adopted an innovative and reliable research approach using two-sample Mendelian randomization to explore the potential causal relationship between insulin use and digestive system malignancies. This is a new attempt to provide us with new ideas for exploring the complex role of insulin and malignant tumors.
Our results showed a positive association between insulin use and CRC, but no significant causal association between insulin use and other GI malignancies. In other words, the use of insulin may reduce the risk of microangiopathy and improve survival in patients with diabetes, but it may increase the risk of colorectal cancer in patients with diabetes. In the study of Chang CH et al. [25], a total of 108,920 patients from the Taiwan National Health Insurance claims database participated in the study, of which 8194 cancer patients were included as the experimental group. 32776 diabetic patients were selected as the control group. Their results showed that the use of any type of insulin increased the risk of cancer (OR, 1.97; 95% CI, 1.85–2.09). In addition, the finding that insulin use was significantly associated with a higher risk of liver, colorectal, lung, gastric, and pancreatic cancer runs counter to our MR Findings. The reason may be related to the confounding factors we mentioned above, for example, the trial was from Taiwan, China, which was an Asian population, while our MR Study selected a European population. In addition, Taiwan is a subtropical island area, and its climate and eating habits are quite different from those of Europe. This can lead to inconsistent results. However, in the study of Gu Y et al. [26], 98 patients in the insulin use group developed cancer, and 170 patients in the non-insulin use group developed cancer. The incidence of cancer in the two groups was 0.786% and 0.743%, respectively. The results showed that there was no significant difference in the risk of cancer between the two groups (RR = 1.20, 95%CI 0.89–1.62, P = 0.228). However, it is thought-provoking that insulin use not only increased the risk of death in insulin users (RR = 1.89, 95%CI 1.47–2.43, P < 0.0001), but also closely related to the increased risk of death in cancer patients (RR = 2.16, 95%CI 1.39–3.35, P = 0.001).This makes it more important for us to develop a set of personalized treatment plans for cancer patients in order to improve the prognosis of patients.
It should be noted that the effect of insulin use on HCC is also highly controversial. Several studies have shown that insulin use is significantly associated with the risk of liver cancer [25, 34], but the results of our study showed no causal association between insulin and liver cancer. We analyzed the inconsistency of our results in two ways. First, our data were from a European database and included a European population, and regional and ethnic differences may cause inconsistent results. Secondly, according to the study results of Yin J et al. [35], chronic hepatitis B OR C infection significantly increased the risk of liver cancer in insulin users (OR = 1.27, 95%CI 1.04–1.55). However, in the Global Burden of hepatitis reports in 2009 and 2019, the incidence of hepatitis B in the European population was significantly lower than that in the Asian population [32, 33]. However, due to the lack of genetic information on insulin users in Asian population, we cannot further investigate the relationship between insulin and HCC in Asian population. Therefore, we recommend that insulin users with chronic viral hepatitis should regularly monitor liver function.
This study has several important advantages.First, the use of TSMR method can avoid the existence of reverse causality in observational studies, and at the same time, the relationship between exposure and outcomes is studied through IVs, which greatly reduces the interference of acquired confounding factors such as social environment and lifestyle on the results. Second, our data were all from the European population and belonged to different columns, which avoided the influence of sample overlap on the results and thus increased the reliability of the MR Results. Finally, the heterogeneity test, pleiotropy test and sensitivity analysis of our results indicated the robustness of our results, and our MR Results were consistent with the results of multiple cohort studies and meta-analyses.
Our study also has some limitations. First, to minimize the effect of population stratification on our results, our data included genetic information only from European populations, which may limit the generalizability of our results. Secondly, because our data came from the aggregated data of GWAS, there was no detailed description of individual data, and the treatment drugs for diabetes patients were often not limited to insulin. Therefore, it is difficult to rule out the bias of the results caused by confounding factors such as age, weight, sex, and medication regimen. Finally, as there is no GWAS data on insulin use in Asia, where the incidence of digestive system cancers is high, more comprehensive GWAS data with larger sample size and further clinical studies are needed to validate our results.