Cadmium and Alzheimer's disease mortality in U.S. adults: Updated evidence with a urinary biomarker and extended follow-up time

Abstract

Cadmium has been linked to impaired cognitive function in adults and may cause behavioral, physiological and molecular abnormalities characteristic of Alzheimer's disease (AD) in animals. Evidence linking cadmium and AD in humans is limited, but supportive. In the most recent epidemiologic study, blood cadmium in U.S. adults was positively associated with elevated AD mortality 7-13 years later. The association between urinary cadmium - an arguably more appropriate biomarker for studying chronic diseases - and AD mortality has not yet been explored. Further study of cadmium and AD mortality in an independent population, with longer follow-up, and stratified by sex is also needed. We sought to answer these questions using the U.S. National Health and Nutrition Examination Survey (NHANES) (1999-2006 cycles) and NHANES III (interviews in 1988-1994) datasets, separately linked to AD mortality as of 2011. We used survey-weighted Cox regression models predicting age at AD death and adjusted for race/ethnicity, sex, smoking status, education and urinary creatinine. An interquartile range (IQR; IQR=0.51ng/mL) increase in urinary cadmium was associated with 58% higher rate of AD mortality (hazard ratio (HR)=1.58, 95% CI: 1.20, 2.09. p-value=0.0009, mean follow-up: 7.5 years) in NHANES 1999-2006 participants. In contrast, in NHANES III participants, an IQR (IQR=0.78ng/mL) increase in urinary cadmium was not associated with AD mortality (HR=0.85, 95% CI: 0.63, 1.17, p-value=0.31, mean follow-up: 13 years). Also in the NHANES III sample however, when the maximum follow-up time was restricted to 12.7 years (i.e. the same as NHANES 1999-2006 participants) and urinary creatinine adjustments were not made, urinary cadmium was associated with elevated AD mortality (HR=1.11, 95% CI: 1.02, 1.20, p-value=0.0086). Our study partially supported an association between cadmium and AD mortality, but the sensitivity of results to follow-up time and creatinine adjustments necessitate cautious interpretation of the association. Further studies, particularly those on toxicological mechanisms, are required to fully understand the nature of the "cadmium-AD mortality" association.

Keywords: Alzheimer's disease; Cadmium; Cognitive function; Heavy metal; NHANES.

Figure 1. Urinary cadmium and predicted log hazard ratio of AD

The figure was based on a survey-weighted Cox regression model where 1) only data in the lower 99% of urinary cadmium was used (N total=1990; N case=21), 2) urinary cadmium was modeled with natural splines with knots at the 25^th, 50^th and 75^th weighted percentiles and 3) the model was adjusted for sex, race/ethnicity, smoking status, education level and urinary creatinine. We used data in the lower 99^th percentile of urinary cadmium because very few observations were present in the upper 1% and the confidence interval for the predicted log HR was large. Including all the data would obscure the presentation of the relationship in the majority (lower 99%) of the data.

Figure 2. Blood cadmium and predicted log hazard ratio of AD

The figure was based on a survey-weighted Cox regression model where 1) only data in the lower 99% of blood cadmium was used (N total=6062; N case=76), 2) blood cadmium was modeled with natural splines with knots at the 10^th, 50^th and 90^th weighted percentiles and 3) the model was adjusted for sex, race/ethnicity, smoking status and education level. We used data in the lower 99^th percentile of blood cadmium because very few observations were present in the upper 1% and the confidence interval for the predicted log HR was large. Including all the data would obscure the presentation of the relationship in the majority (lower 99%) of the data.