Introduction:
Low-moderate exposure to inorganic arsenic has been associated with hemoglobin A1c (HbA1c) in some studies. While HbA1c is a marker of diabetes control among individuals with treated diabetes, the association between arsenic and HbA1c has not been evaluated according to diabetes status. Hypothesis: Arsenic is positively associated with HbA1c. Methods: We included 3,567 (1774 with diabetes and 1793 without diabetes) American Indian adults who participated in the Strong Heart Study in 1989-1991 with measured urine arsenic, HbA1c and other covariates. We compared HbA1c levels by urine arsenic quartiles using linear regression stratified by diabetes status and adjusted for urine creatinine, demographic variables, body mass index and smoking. Among diabetics, we ran models stratified by fasting glucose and by diabetes treatment. Results: Median (interquartile range) HbA1c levels were 8.1 (6.1, 10.3)% and 5.1 (4.7, 5.4)%, in participants with and without diabetes, respectively. Adjusted mean HbA1c levels were 0.55 (95% CI 0.20, 0.90)%, 1.13 (0.76, 1.49)% and 1.28 (0.77, 1.59)% higher in quartiles 2-4 compared to the lowest quartile; p for trend <0.001. Urine arsenic was not associated with HbA1c levels in those without diabetes: mean difference in HbA1c levels comparing quartiles 2-4 to the lowest quartile were 0.08 (0.01, 0.15)%, -0.03 (-0.10, 0.05)% and -0.05 (-0.13, 0.04)%, respectively; p for trend =0.10. Among diabetics, the association between arsenic and HbA1c was strongest among participants with fasting glucose ≥180 mg/dL, in untreated participants, and in those taking only oral hypoglycemics (Table).
Difference (95% Confidence Interval) in HbA1c by Total Urine Arsenic (adjusted for age, sex, region, obesity, smoking and urine creatinine)
Fasting Plasma Glucose (mg/dL)
Diabetes Treatment
Urine Arsenic (μg/L)
<1.6 (N=291)
≥126 and <180 (N=594)
≥180(N=865)
None (N=688)
Oral(N=637)
Insulin (N=436)
Q1 (<8.83)
0.00 (referent)
0.00 (referent)
0.00 (referent)
0.00 (referent)
0.00 (referent)
0.00 (referent)
Q2 (≥8.83 and <15.38)
0.15 (-0.33,0.64)
0.04 (-0.34, 0.42)
0.41 (0.03, 0.80)
0.57 (0.07, 1.06)
0.37 (-0.15, 0.89)
0.35 (-0.27, 0.97)
Q3 (≥15.38 and <26.37)
0.43 (-0.12, 0.98)
0.38 (-0.04, 0.80)
0.49 (0.09, 0.90)
0.93 (0.39, 1.46)
0.73 (0.19, 1.28)
0.61 (-0.07, 1.28)
Q4(≥26.37)
−0.21 (-0.86, 0.44)
0.27 (-0.17, 0.71)
0.58 (0.12, 1.04)
0.91 (0.34, 1.49)
0.90 (0.25, 1.54)
0.36 (-0.38, 1.11)
p for trend
*
0.29
0.22
0.08
0.02
0.01
0.69
*
Based on quartile medians Urine arsenic concentrations were positively associated with HbA1c in Strong Heart Study participants with diabetes. No association was found among participants without diabetes. Prospective analyses that evaluate the relationship between arsenic exposure and diabetes, including measures of diabetes control, are needed.