Diabetes Correlates With Increased Risk of Pancreatic Cancer
Diabetes Correlates With Increased Risk of Pancreatic Cancer
This study consisted of 49 803 subjects in the diabetic group and 199 212 subjects in the non-diabetic group, with similar sex and age distributions and mean age of 55.9 years (Table 1). The diabetic group had higher prevalence of obesity, chronic pancreatitis, alcoholism, gallstones, hepatitis B and hepatitis C infections at the baseline (P < 0.001).
During the follow-up period, the incidence of pancreatic cancer was 1.79-fold greater in the diabetic group than in the non-diabetic group (1.78 vs 1.00 per 10 000 person-years, 95%CI = 1.53–2.10) (Table 2). This further translated into approximately eight additional cases of pancreatic cancer per 100 000 diabetic patients each year. Regardless of sex or age groups, the incidence of pancreatic cancer was still higher in subjects with DM. The incidence of pancreatic cancer was higher in patients with diabetic duration less than 2 years, as compared to the non-diabetic group (27.81 vs 6.96 per 10 000 person-years, 95%CI = 2.11–7.54). After 2 years of diabetes diagnosis, however, there was no statistical significance in the incidence between the diabetic group and the non-diabetic group (1.21 vs 0.86 per 10 000 person-years, 95%CI = 0.96–2.03).
Adjusted hazard ratios and 95%CI of pancreatic cancer associated with DM and comorbidities are shown in Table 3. The adjusted HR of pancreatic cancer for patients with DM was 1.77 (95%CI = 1.29–2.43). Chronic pancreatitis (HR = 19.40, 95%CI = 10.36–36.30), gallstones (HR = 2.56, 95%CI = 1.71–3.82), and hepatitis C infection (HR = 3.08, 95%CI = 1.74–5.44) were also significant factors predicting pancreatic cancer. When compared with subjects aged 20–39, subjects aged 40–64 and subjects aged 65 and older also had higher HR of pancreatic cancer (HR = 5.22, 95%CI = 1.29–21.17 and HR = 7.59, 95%CI = 1.86–31.00, respectively).
The case number was too small for hepatitis C infection. Therefore, we combined the status of hepatitis C infection together. Table 4 shows stratified analyses by the status of DM, chronic pancreatitis, gallstones and/or hepatitis C infection for hazards associated with pancreatic cancer controlling for age. There were synergistic effects on the pancreatic cancer hazards between factors. Subjects comorbid with DM and chronic pancreatitis had the highest HR of pancreatic cancer, as compared with subjects without these comorbidities (HR = 33.52, 95%CI = 10.61–105.94).
Table 5 shows the effects of anti-diabetic drugs on the risk of pancreatic cancer. After adjustments for potential confounders, there was no significant association between use of anti-diabetic drugs and the risk of pancreatic cancer.
Results
Baseline Characteristics of the Study Population
This study consisted of 49 803 subjects in the diabetic group and 199 212 subjects in the non-diabetic group, with similar sex and age distributions and mean age of 55.9 years (Table 1). The diabetic group had higher prevalence of obesity, chronic pancreatitis, alcoholism, gallstones, hepatitis B and hepatitis C infections at the baseline (P < 0.001).
During the follow-up period, the incidence of pancreatic cancer was 1.79-fold greater in the diabetic group than in the non-diabetic group (1.78 vs 1.00 per 10 000 person-years, 95%CI = 1.53–2.10) (Table 2). This further translated into approximately eight additional cases of pancreatic cancer per 100 000 diabetic patients each year. Regardless of sex or age groups, the incidence of pancreatic cancer was still higher in subjects with DM. The incidence of pancreatic cancer was higher in patients with diabetic duration less than 2 years, as compared to the non-diabetic group (27.81 vs 6.96 per 10 000 person-years, 95%CI = 2.11–7.54). After 2 years of diabetes diagnosis, however, there was no statistical significance in the incidence between the diabetic group and the non-diabetic group (1.21 vs 0.86 per 10 000 person-years, 95%CI = 0.96–2.03).
Pancreatic Cancer and Comorbidities by Univariate and Multivariate Cox Proportional Hazard Analysis
Adjusted hazard ratios and 95%CI of pancreatic cancer associated with DM and comorbidities are shown in Table 3. The adjusted HR of pancreatic cancer for patients with DM was 1.77 (95%CI = 1.29–2.43). Chronic pancreatitis (HR = 19.40, 95%CI = 10.36–36.30), gallstones (HR = 2.56, 95%CI = 1.71–3.82), and hepatitis C infection (HR = 3.08, 95%CI = 1.74–5.44) were also significant factors predicting pancreatic cancer. When compared with subjects aged 20–39, subjects aged 40–64 and subjects aged 65 and older also had higher HR of pancreatic cancer (HR = 5.22, 95%CI = 1.29–21.17 and HR = 7.59, 95%CI = 1.86–31.00, respectively).
Interaction between Diabetes and Comorbidities
The case number was too small for hepatitis C infection. Therefore, we combined the status of hepatitis C infection together. Table 4 shows stratified analyses by the status of DM, chronic pancreatitis, gallstones and/or hepatitis C infection for hazards associated with pancreatic cancer controlling for age. There were synergistic effects on the pancreatic cancer hazards between factors. Subjects comorbid with DM and chronic pancreatitis had the highest HR of pancreatic cancer, as compared with subjects without these comorbidities (HR = 33.52, 95%CI = 10.61–105.94).
Influence of anti-diabetic Drugs on the Risk of Pancreatic Cancer
Table 5 shows the effects of anti-diabetic drugs on the risk of pancreatic cancer. After adjustments for potential confounders, there was no significant association between use of anti-diabetic drugs and the risk of pancreatic cancer.