People in southern Finland who drank water from drilled bedrock wells, some with very high natural uranium, gave water, urine and blood samples. Urinary uranium was linked to higher fractional excretion of calcium and phosphate, a sign of altered proximal tubule function, with no clear threshold. Glomerular measures such as creatinine clearance and urinary albumin were not associated with exposure.
Key numbers
- Uranium in drinking water: median 28 µg/L (interquartile range 6 to 135; maximum 1,920 µg/L), n = 325 (abstract, p. 337); full range 0.001 to 1,920 µg/L, with 30% above 100 µg/L (p. 340).
- Uranium in urine: median 13 ng/mmol creatinine (interquartile range 2 to 75) (abstract, p. 337).
- Daily uranium intake from drinking water: median 39 µg (interquartile range 7 to 224) (abstract, p. 337; Table 3).
- Per 1 µg/mmol creatinine increase in urinary uranium: fractional calcium excretion +1.5% (95% CI 0.6 to 2.3), phosphate +13% (1.4 to 25), glucose excretion +0.7 µmol/min (-0.4 to 1.8) (abstract, p. 337; Table 4).
- Phosphate fractional excretion was significantly higher at water uranium above 300 µg/L than below 2 µg/L (p. 340; Table 5).
- Detection limits: 0.0004 µg/L for uranium in water and 0.002 µg/L in urine by ICP-MS (Table 2, p. 338).
- Background cited by the authors: wells in Finland with up to 12,000 µg/L, a WHO proposed guideline of 2 µg/L at that time, and about one-seventh of the Finnish population using water above 2 µg/L (p. 337).
Methods (brief)
Cross-sectional study. Uranium in water and overnight urine was measured blind by ICP-MS with split-sample and inter-laboratory checks; kidney markers were measured in a hospital clinical chemistry laboratory. Generalized linear models were adjusted for age, sex and body mass index. The study was funded by the Ministry of Social Affairs and Health of Finland.
Evidence fitness
Supports statements that chronic uranium intake from drinking water is weakly associated with proximal tubule changes in humans, and gives measured water, urine and intake distributions. It does not show clinical kidney disease or glomerular damage.
Limitations
Cross-sectional design, a single urine collection, modest effect sizes of uncertain clinical meaning, and exclusion of pregnant women and people with diabetes.
Related evidence
Update history
No substantive edit history is available in this build. The full commit record is available in git.