| Abstract: | Background: Whether insulin use predicts bladder cancer mortality has not been investigated. Furthermore, it is not known whether insulin use and smoking jointly influence the risk. Materials and Methods: A total of 86,939 patients (40,014 men, 46,925 women) with type 2 diabetes mellitus (T2DM) and aged ≥ 25 years in a nationally representative cohort were followed prospectively from 1995 to 2006 for bladder cancer mortality. Cox regression was used considering the following independent variables: age, sex, diabetes duration, body mass index, smoking, insulin use, and area of residence. The models were created for patients aged ≥ 25 and ≥ 65 years, separately; and sensitivity analyses were conducted after excluding (1) patients with duration between onset of diabetes and bladder cancer mortality < 5 years, and (2) patients with diabetes duration at recruitment < 3 years. The joint effect of insulin use and smoking was evaluated. Results: Patients who died of bladder cancer were characterized by older age, male predominance, longer diabetes duration, smoking, and insulin use. In multivariable Cox models, age, male sex, and insulin use were consistently predictive for bladder cancer mortality in all analyses, whereas the other variables were not. The adjusted hazard ratios for bladder cancer mortality for insulin users vs. nonusers ranged from 1.877 to 2.502 in different models (all P values < .05). Insulin use and smoking jointly increased the adjusted hazard ratio to 3.120 (95% confidence interval, 1.329-7.322). Conclusions: Insulin use is significantly predictive for bladder cancer mortality in patients with T2DM. Insulin use and smoking jointly increase the risk. |
