Table 5.

Associations between prevalence of diabetes and quartiles of serum concentrations of dioxin-like PCB TEFs and non-dioxin-like, non- and mono-ortho PCBs.

Variables	Quartile 1	Quartile 2	Quartile 3	Quartile 4	Wald χ2 p-value
A. Congeners with values below the MDL given concentrations of the MDL/square root of 2.
Dioxin-like PCBs TEQ
N (diabetics)	145 (12)	163 (13)	49 (28)	144 (58)
Model 1
OR (95% CI)	1.0 (ref)	0.61 (0.25–1.46)	1.46 (0.64–3.32)	3.08 (1.27–7.49)
p-Value	—	0.27	0.37	0.013	0.0004
Model 2
OR (95% CI)	1.0 (ref)	0.49 (0.18–1.34)	0.93 (0.33–2.59)	1.82 (0.61–5.40)
p-Value	—	0.17	0.88	0.28	0.010
Non-dioxin-like PCBs
Model 1
N (diabetics)	155 (7)	149 (18)	150 (32)	147 (54)
OR (95% CI)	1.0 (ref)	1.87 (0.72–4.80)	3.33 (1.32–8.40)	6.01 (2.32–15.59)
p-Value	—	0.20	0.011	0.0002	0.0003
Model 2
OR (95% CI)	1.0 (ref)	2.00 (0.72–5.58)	2.92 (1.07–7.96)	5.01 (1.76–14.24)
p-Value	—	0.18	0.037	0.0025	0.0088
B. Congeners with more than 40% of concentrations below the MDL removed from the analysis.
Non-dioxin-like PCBs
N (diabetics)	155 (10)	155 (15)	144 (27)	147 (59)
Model 1
OR (95% CI)	1.0 (ref)	1.13 (0.47–2.69)	1.86 (0.79-4.40)	4.61 (1.87–11.33)
p-Value	—	0.79	0.16	0.0009	0.0002
Model 2
OR (95% CI)	1.0 (ref)	1.10 (0.42–2.85)	1.49 (0.56–3.98)	3.53 (1.26–9.89)
p-Value	—	0.85	0.43	0.016	0.0077
Note: The number (N) of subjects in each quartile is given and the number of subjects with diabetes is given in parentheses. Model 1–results are adjusted for age, gender, BMI, and serum concentrations of total lipids, but not for total pesticides. Model 2–results are adjusted for total pesticides as well as age, gender, BMI, serum concentrations of total lipids. Wald χ2 p-value is the significance level of the differences in proportions of diabetes across the quartiles of each predictor variable.