An etiologic prediction model incorporating biomarkers to predict the bladder cancer risk associated with occupational exposure to aromatic amines: a pilot study

Table 2 Logistic regression (full and short models)

Variables	Classes	Full model			Short model
Variables	Classes	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
Smoking^b	Ex-smokers	0.05	0.00–1.47	0.083	0.11	0.01–1.49	0.096
Smoking^b	Smokers	1.16	0.16–8.60	0.887	1.56	0.31–7.94	0.589
Coffee consumption^c	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 adducts^d	≥1	19.20	2.52–146.	0.004	6.02	1.66–21.8	0.006
Genotypes^e	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
Pearson chi-square test:	no. covariates		197		44
Hosmer-Lemeshow test:	p-value		0.130		0.942
Hosmer-Lemeshow test:	no. groups		10		10

ISSN: 1745-6673