Urinary excretion of cadmium among Torres Strait Islanders (Australia) at risk of elevated dietary exposure through traditional foods

This study explored urinary cadmium levels among Torres Strait Islanders in response to concerns about potential health impact of high levels of cadmium in some traditional seafood (dugong and turtle liver and kidney). Cadmium levels were measured by inductively coupled mass spectrometry in de-identified urine samples collected during general screening programs in 1996 in two communities with varying dugong and turtle catch statistics. Statistical analysis was performed to identify links between cadmium levels and demographic and background health information. Geometric mean cadmium level among the sample group was 0.83 mu g/g creatinine with 12% containing over 2 mu g/g creatinine. Cadmium level was most strongly associated with age (46% of variation), followed by sex (females > males, 7%) and current smoking status (smokers > non-smokers, 4.7%). Adjusting model conditions suggested further positive associations between cadmium level and diabetes (p = 0.05) and residence in the predicted higher exposure community (p = 0.07). Positive correlations between cadmium and body fat in bivariate analysis were eliminated by control for age and sex. This study found only suggestive differences in cadmium levels between two communities with predicted variation in exposure from traditional foods. However, the data indicate that factors linked with higher cadmium accumulation overlap with those of renal disease risk (i.e. older, females, smokers, diabetes) and suggest that levels may be sufficient to contribute to renal pathology. More direct assessment of exposure and health risks of cadmium to Torres Strait Islanders is needed given the disproportionate level of diet-related disease and the cultural importance of dugong and turtle. This study highlights the need to consider social and cultural variation in exposure and to de. ne "safe'' cadmium levels during diabetes given its rising global prevalence.