Table 2.

Logistic regression (full and short models)

Variables	Classes	Full model			Short model
		OR	95% CI	p-value	OR	95% CI	p-value
Age (years)a	57–65.9	0.12	0.01–0.98	0.048
	66–70.9	0.29	0.04–2.25	0.239
	≥ 71	0.10	0.01–1.07	0.056
Smokingb	Ex-smokers	0.05	0.00–1.47	0.083	0.11	0.01–1.49	0.096
	Smokers	1.16	0.16–8.60	0.887	1.56	0.31–7.94	0.589
Coffee consumptionc	2	0.63	0.09–4.64	0.654
	3	1.96	0.27–14.24	0.505
	≥4	0.05	0.00–0.86	0.039
DNA adductsd	≥1	19.20	2.52–146.	0.004	6.02	1.66–21.8	0.006
Genotypese	GSTM1	3.01	0.51–17.7	0.223
	GSTT1	0.29	0.04–1.92	0.198
	GSTP1	0.19	0.03–1.08	0.061	0.41	0.12–1.38	0.150
	NAT1	0.27	0.05–1.52	0.139
	NAT2	2.55	0.52–12.6	0.25
	MPO	1.38	0.03–65.1	0.871
	COMT	0.05	0.01–0.39	0.005	0.21	0.06–0.72	0.014
	MnSOD	0.30	0.06–1.49	0.142
	CYP1B1	0.27	0.03–2.42	0.24
	XRCC1	0.97	0.22–4.22	0.966
	XRCC3	0.38	0.09–1.68	0.203	0.40	0.12–1.34	0.139
	XPD	0.31	0.05–1.81	0.195
Constant term		5.95	0.06–545.	0.439	0.18	0.03–1.30	0.09
Reference groups: apatients with ≤56.9 years; bnon smokers and ex-smokers from >20 years; cpatients with consumption of ≤1 cup/day; dpatients with <1 (logarithm values) of DNA adducts×108 nucleotides; epatients with genotype values set at 0 (see text for definition).
Measures of fit for logistic regression			Full model		Short model
R-square			0.389		0.228
Area under the ROC curve			0.931		0.856
Pearson chi-square test:	p-value		0.087		0.826
	no. covariates		197		44
Hosmer-Lemeshow test:	p-value		0.130		0.942
	no. groups		10		10