Approaches for Analyzing the Temporal Relationship Between Arsenic Exposure and Bladder Cancer Over the Life Course

Abstract

P-729 Introduction: Ingestion of arsenic in drinking water has been linked with the risk of urinary bladder cancer, however, questions remain about the characterization of the dose-response relationships, ranges in induction-latency intervals, and whether there are critical lifetime exposure periods of heightened susceptibility. In Michigan, approximately 235,000 people are exposed to arsenic concentrations above the new maximum contaminant limit (10 μg/L) in their home drinking water. The risks from these moderately elevated arsenic concentrations are being investigated in a population-based bladder cancer case-control study in Michigan (220 cases and 440 controls to date). Methods: Water samples were collected and analyzed for arsenic at current residences, and inorganic arsenic concentrations were estimated at past residences. Questionnaires were used to obtain residential history, dietary habits, and smoking history from participants. Individual-level lifetime estimates of arsenic exposure were calculated using STIS (TerraSeer, Inc., Crystal lake, IL), integrating sources of spatial, temporal, and spatio-temporal variability. An approach for analyzing yearly estimates of arsenic exposure is introduced, and compared with more traditional metrics such as cumulative, current, and decade-long time windows of exposure. Unconditional logistic regression analyses were conducted using “proc logistic” in SAS (SAS Institute, Inc., Cary, NC); both participant's age and calendar year were considered in the timing of exposure. Results: At the time of this analysis, data collection is incomplete and no inferences about exposure-disease relationships should be drawn – these data are analyzed for the sole purpose of demonstrating these novel methods. When comparing those exposed to arsenic > 10 μg/day with those exposed to < 5 μg/day in models adjusted for race, gender, age, education, and smoking, no relationship was observed between arsenic exposure and bladder cancer when using current (OR=1.19; 95% CI: 0.71, 2.00), cumulative (OR=1.00; 95% CI: 0.61, 1.65), or decade-long time windows of exposure (ORs ranged from 0.72–1.48; lower end of 95% CIs 0.40–0.92; upper end CIs 1.27–2.60, for different decades). Yearly estimates of exposure, however, revealed an association between arsenic exposure and bladder cancer, if exposure occurred when participants were in their 40s (ORs ranged from 1.80–1.89; lower end of 95% CI 1.05–1.13; upper end CI 3.10–3.52); analyses based on calendar year were not significant. Discussion and Conclusions: Although significant results from analyses of yearly exposure estimates must be interpreted cautiously because of potential spurious results that may accompany the large number of statistical analyses, this approach has the potential to generate valuable insights for characterizing the temporal relationship between exposure and disease.