Several epidemiological studies investigated diabetes as a risk factor for bladder cancer. A meta-analysis of 16 studies reported a relative risk of bladder cancer of 1.24 in adults with diabetes. However, many of those studies did not adjust for potential confounders. The Iowa Women's Health Study (IWHS) was included in that meta-analysis, but during the follow-up (1986–1998), there were only 112 incident bladder cancer cases, 12 of which had diabetes reported at baseline.

We updated the analysis of the association between diabetes and bladder cancer occurrence (n=265) in the IWHS with follow-up through 2008. Our goal was to more accurately estimate the risk of bladder cancer in relation to diabetes at baseline and during follow-up: 1987, 1989, 1992, and 1997. We also studied whether the association varied by the time between the self-reported onset of diabetes and bladder cancer (diabetes duration), and by obesity measures – body mass index (BMI) and waist-to-hip ratio (WHR). We used Cox proportional hazards regression to estimate hazard ratio (HR) of bladder cancer and 95% confidence interval (CI) for baseline diabetes, and extended Cox regression to conduct a time-dependent analysis for diabetes during follow-up. To examine an association of diabetes duration with bladder cancer incidence, we conducted a nested case-control analysis. For each case, 4 controls were randomly chosen among all IWHS participants who were alive and free of cancer on the date of cancer diagnosis. Controls were matched to cases by year of birth and were assigned an index date, equivalent to the diagnosis date of the matched case. Conditional logistic regression was used to estimate the odds ratio (OR) of bladder cancer.

The cohort for the person-time analysis included 37,459 initially cancer-free post-menopausal women (mean age at baseline was 61.7 y, 99.9% were white). In the multivariate analysis diabetics at baseline had an HR=1.51 (95% CI, 0.91;2.50) compared to non-diabetics. The weaker association in the present than in an earlier analysis (HR=2.5, 95% CI, 1.3;4.5) could be due to a prolonged follow-up and non-differential misclassification of diabetes status and/or survival bias.

In the time-dependent analysis accounting for incident diabetes after the baseline, the HR for bladder cancer was 1.77 (95% CI, 1.24;2.51) for yes versus no diabetes. The association did not markedly change after excluding women with an onset of diabetes before 30 years of age or with diabetes duration <2 y. There was no interaction between diabetes and smoking. After stratification by BMI, HRs of bladder cancer for diabetics versus non-diabetics were 1.29 (95%CI, 0.60;2.79), 1.52 (95%CI, 0.83;2.79), and 2.60 (95%CI, 1.46;4.61) for BMI<25, 25–30 and >30 kg/m2, respectively. Associations between diabetes and bladder cancer were stronger for higher WHR. Finally, in the nested case-control analysis, the HRs of bladder cancer for patients with diabetes duration <5, 5–10, and >10 y were 1.65 (95%CI, 0.76; 3.60), 2.18 (95%CI, 0.85;5.60), 3.37 (95%CI, 0.88;12.87) (p-trend=0.03).

In summary, we have confirmed a statistically significant positive association between diabetes and bladder cancer risk among white post-menopausal women. It may be explained by the mitogenic properties of insulin and/or increased likelihood of urinary tract infections in diabetics. The association was stronger with longer diabetes duration and with increasing adiposity. The latter association may reflect more severe diabetes among obese individuals and its larger influence on bladder carcinogenesis.

Citation Information: Cancer Prev Res 2011;4(10 Suppl):A76.