The relationship between dietary antioxidants and oxidative damage in smokers: Evidence of effect modification by lifestyle and genetic factors.

PR-05

Several lines of evidence point to a role for vitamins in cancer prevention, though conflicting results indicate a need for more sophisticated analyses. To better understand the possible role of antioxidants in the chemoprevention of tobacco-related cancer, we measured plasma α-tocopherol levels and 8-hydroxy-2'-deoxyguanosine (8OHdG) in WBC DNA of 280 smokers prior to treatment as part of a randomized clinical trial of antioxidant supplementation. We evaluated the main effects of α-tocopherol on oxidative damage (8OHdG) in WBC DNA and possible effect modification by gender, obesity, GSTM1 polymorphism, and amount smoked. Adjusted and crude models were not meaningfully different, so crude results are reported. As previously reported in this study, there was a protective effect of plasma antioxidants (α-tocopherol) on oxidative damage among smokers (β = -.12, p=0.03) (Garduno et al AACR abstract 2004), restricted to a protective effect among men, with no apparent effect among women (β (men)= -0.223, p<0.01, β (women)= 0.025, p=0.78) (Madsen, et al AACR abstract 2005). In the current analyses, increasing plasma α -tocopherol was associated with reduced oxidative damage in non-obese subjects (βunadjusted = -0.17, p =<0.01) but not in obese subjects (βunadjusted = 0.13, p =0.30). The interaction term was significant (βobese x α -tocopherol= .0.30, p=0.04). Increasing plasma α -tocopherol was associated with reduced oxidative damage in GSTM1 non-null subjects (βunadjusted = -0.19, p =<0.01)) but not in GSTM1 null subjects (βunadjusted = - 0.005, p =0.95). The interaction term was suggestive (βGSTM1 x α -tocopherol= 0.18, p=0.10). Plasma α -tocopherol was associated with reduced oxidative damage in light smokers (βunadjusted = -0.19, p =<0.01) but not in heavy smokers (βunadjusted = 0.05, p = 0.50). The interaction term was significant (βheavy smoking x α -tocopherol = 0.24, p= 0.05). The final model considered the main effects of gender, obesity, GSTM1, amount smoked, and two-way interaction terms for each variable with plasma α-tocopherol on 8OHdG. Plasma α-tocopherol, gender, and obesity main effects were significant at p<0.05 and the p-values for plasma-α-tocopherol x gender, plasma-α-tocopherol x obesity, plasma α-tocopherol x GSTM1, plasma α-tocopherol x heavy smoking, and heavy smoking terms were between p=0.05 and p=0.10. Our analyses suggest some potentially important interactions governing any protective effect of antioxidants against smoking-related oxidative damage.