Associations of drinking water and urinary uranium with incident diabetes among U.S. adults: The Multi-Ethnic Study of Atherosclerosis and the Strong Heart Family Study.

OBJECTIVES

We examined the association of uranium in community water systems (CWS) and in urine with type 2 diabetes (T2D) incidence in the Strong Heart Family Study (SHFS), a prospective study of American Indian communities, and the Multi-Ethnic Study of Atherosclerosis (MESA), a prospective study of racially and ethnically diverse urban U.S. communities.

METHODS

We evaluated participants from SHFS (N = 1609) and MESA (N = 3080) with CWS and urine uranium available, free of T2D at baseline (2001-2003 and 2000-2002, respectively). We used Cox mixed-effects models to account for clustering by residential ZIP Code and family (SHFS), and estimated hazard ratios (HRs) of incident T2D per log2-increase and across categories of CWS and urine uranium adjusted for T2D risk factors and arsenic levels. Urinary models accounted for kidney function.

RESULTS

Incident T2D cases were 198 in SHFS (mean follow-up 5.6 years) and 525 in MESA (mean follow-up 13.8 years). In a meta-analysis across SHFS and MESA, the adjusted HR (95% CI) of incident T2D was 0.97 (0.89, 1.04) and 1.01 (0.97, 1.06) per log-doubling of CWS and urine uranium, respectively. We observed a stronger association between uranium exposure and T2D among participants with lower BMI compared to higher BMI in both cohorts, although this finding was limited by small case sizes in the lowest BMI subgroups.

CONCLUSIONS

CWS and urine uranium were not statistically associated with T2D across U.S. cohorts with low uranium exposure (median <1.2 μg/L). Given the importance of uranium exposure, further evidence is needed at low-to moderate-exposure levels.