Residential proximity to dioxin-emitting facilities and risk of non-Hodgkin lymphoma in the NIH-AARP Diet and Health Study

Background: Polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/Fs) are persistent organic pollutants emitted from industrial sources. Residential proximity to these emissions has been associated with risk of nonHodgkin lymphoma (NHL) in a limited number of studies. Methods: We evaluated associations between residential proximity to PCDD/F-emitting facilities and NHL in the NIH-AARP Diet and Health Study (N = 451,410), a prospective cohort enrolled in 1995-1996 in 6 states and 2 U. S. cities. We linked enrollment addresses with a U.S. Environmental Protection Agency database of 4,478 historical PCDD/F sources with estimated toxic equivalency quotient (TEQ) emissions. We evaluated associations between NHL and exposures during a historical period prior to enrollment (1980-1995) using an average emissions index, weighted by toxicity, distance, and wind direction (AEI -W [g TEQ/km 2 ]) within 3-, 5- and 10 km of residences. We also evaluated proximity -only metrics indicating the presence/absence of one or more facilities within each distance, and metrics calculated separately for each facility type. We used Cox regression to estimate associations (hazard ratio, HR; 95 % confidence interval, 95 %CI) with NHL and major subtypes, adjusting for demographic, lifestyle, and dietary factors. Results: A total of 6,467 incident cases of NHL were diagnosed through 2011. Participants with an AEI -W >= 95th percentile had elevated risk of NHL compared to those unexposed at 3 km (HR = 1.16; 95 %CI = 0.89-1.52; ptrend = 0.24), 5 km (HR = 1.20;95 %CI = 0.99-1.46;p -trend = 0.05) and 10 km (HR = 1.15; 95 %CI = 0.99-1.34; p -trend = 0.04). We found a positive association at 5 km with follicular lymphoma (HR >= 95vs.0 = 1.62; 95 %CI = 0.98-2.67; p -trend = 0.05) and a suggestive association for diffuse large B -cell lymphoma (HR >= 95vs.0 = 1.40; 95 %CI = 0.91-2.14; p -trend = 0.11). NHL risk was also associated with high emissions from coal-fired power plants within 10 km (HR >= 95vs.0 = 1.42; 95 %CI = 1.09-1.84; p -trend = 0.05). Conclusions: Residential proximity to relatively high dioxin emissions from industrial sources may increase the risk of NHL and specific subtypes.