A187

Our objective was to test the hypothesis that in pancreatic ductal adenocarcinoma (PDA) mechanistic causal relationships exist among concentrations of organochlorine compounds (OC) and the occurrence or persistence of K-ras mutations, as well as among the latter and coffee intake. This develops simpler analyses based on smaller case-case designs (Lancet 1999, JECH 1999). Incident cases of PDA were interviewed and had blood drawn soon during hospital admission (N = 103 patients with information on K-ras, OC and interview). OC were measured by high-resolution gas chromatography with electron-capture detection. ORs and 95%CI were computed by exact logistic regression. Exposure-response relationships were also assessed with smoothing splines. Even after adjusting by p,p'-DDT, mutations were more likely among regular coffee drinkers (RCD) than among non-RCD (OR: 4.2; 95%CI: 1.1-5.5; p=0.031). As compared to non-RCD, the OR of a mutated tumour, adjusted by age, sex and p,p'-DDE, was 4.2 for drinkers of 1-7 cups/week (c/w), 5.2 for 8-14 c/w and 6.1 for 15+ c/w (p-trend=0.022). ORs were similar when adjusting by other OC and by cancer symptoms; eg, when adjusting by age, sex, PCB 153 and cholestatic signs, the ORs (coffee c/w) were 4.1, 8.4 and 8.0 (p-trend=0.014). The OR for p,p'-DDT adjusted by age, sex and coffee was 7.9 (95%CI: 2.5-25.2; p<0.001); for subjects in mid and upper tertiles of PCB 138 the OR were 2.8 and 6.6 (p-trend=0.008). The associations were stronger for the 2 most prevalent mutations (Arg and Val); eg, for mid and upper tertiles of p,p'-DDE the OR for Val were 3.7 and 9.3 (p-trend=0.033). OC and coffee were weakly or not correlated (Spearman's rho always <0.1 and p >0.4). No significant interactions were apparent among OC and coffee intake (p>0.67 always). Upon adjustment by age, sex, symptoms, coffee and 2 OC, p,p'-DDT (p=0.029), PCB 153 (p=0.025) and coffee (p=0.021) all remained strongly associated with the mutation. The association was not indiscriminate with all OC: concentrations of HCB and β-HCH were also high, and yet these OCs were not associated with an increased probability of a K-ras mutated tumour. Some OC and coffee may have a role in the etiopathogenesis of exocrine pancreatic cancer through modulation of K-ras activation or persistence. Results are coherent with mechanistic hypotheses on a nongenotoxic, indirectly genotoxic or epigenetic role of some OC and of coffee. The high prevalence in human cancers of these acquired genetic alterations in the K-ras oncogene, and the generalized human contamination by OC make it especially relevant to refute or to replicate the findings.

[Fifth AACR International Conference on Frontiers in Cancer Prevention Research, Nov 12-15, 2006]