TABLE 5. Multivariable ordinal logistic regression analysis comparing participants with prediabetes^a and diabetes^b based on urinary arsenic concentrations quartiles, CHMS, Cycle 1, 2007–2009.

Urinary arsenic, µg/Lc	Number of participantsd			OR (95% CI)
	Controls (n = 2054)	With prediabetesa (n = 831)	With diabetesb (n = 225)	Crude OR	Adjusted OR (Model 1)e	Adjusted OR (Model 2)f
< 5.71	554	171	46	1.00 (Referent)	1.00 (Referent)	1.00 (Referent)
5.71–11.20	520	197	54	1.20 (0.98–1.47)	1.35 (0.95–1.79)	1.35 (0.97–1.82)
11.21–22.98	530	192	64	1.20 (0.88–1.64)	1.39 (1.01–2.00)	1.41 (1.02–2.04)
≥ 22.99	450	271	61	1.56 (1.00–2.44)	1.85 (1.11–3.13)	1.89 (1.12–3.13)
p for trend				.049	.019	.016

Abbreviations: BMI, body mass index; CHMS, Canadian Health Measures Survey; CI, confidence interval; HbA1c, glycated hemoglobin; OR, odds ratio.

^aFasting glucose 100–125 mg/dl (5.6–6.9 mmol/L) or HbA1c 5.7%–6.4%.

^bFasting glucose ≥ 126 mg/dL or HbA1c ≥ 6.5% or self-reported medication use or self-reported health care professional diagnosis.

^cUrinary arsenic not corrected for urinary creatinine.

^dData missing for n = 41 participants.

^eModel 1 adjusted for urinary creatinine, age, sex, alcohol status, smoking status, educational status, BMI and hypertension.

^fModel 2 adjusted as for Model 1 plus for seafood consumption.