Background: We extend our work studying environmental and genetic determinants of polycyclic aromatic hydrocarbon (PAH) DNA adducts in radical prostatectomy specimens, to consider cross-level interactions between cigarette smoking and indicators of neighborhood level socio-economic status.
Methods: PAH-DNA adducts were measured in 397 prostatectomy specimens from the Henry Ford Health System using immunohistochemistry with image analysis to measure staining intensity in optical density units. Subjects' home addresses were geo-coded to Census tracts and linked to 2000 Census data. Tracts were classified for educational attainment using the median value across tracts for the percentage of residents who graduated college. GEE models, accounting for clustering at the Census tract level, were used to determine if smoking was associated with adduct levels in tumor tissue by strata of neighborhood educational attainment. Analyses adjusted for race, age, tumor volume, primary Gleason grade and PSA level at diagnosis.
Results: Among those living in tracts with high educational attainment, smoking status predicted adduct levels. The covariate adjusted mean staining intensity for current smokers was 0.17 (95% CI = 0.15-0.19), for ex-smokers was 0.16 (95% CI = 0.15-0.17) and never-smokers was 0.13 (95% CI = 0.12-0.14). For those living in tracts with low educational attainment there was no significant difference in adduct levels by smoking status, the covariate adjusted mean staining intensity for current smokers was 0.16 (95% CI = 0.14-0.18), for ex-smokers was 0.15 (95% CI = 0.14-0.16) and for never smokers was 0.16 (95% CI = 0.15-0.17). The P-value for the interaction term between smoking status and tract level educational attainment was 0.02. Further adjustment for individual level education and for tract median household income did not alter these results.
Conclusion: The results suggest that neighborhood context modifies the relationship between individual smoking status and PAH-DNA adduct levels in prostate tissue; smoking is only predictive of adduct levels in higher SES tracts. The spatial segregation of income groups in and around Detroit suggests that indicators of lower neighborhood SES serve as a proxy for other environmental sources of PAH.