Observational data have suggested that increased water intake decreases the risk of CHD. A postulated mechanism is that increased water ingestion reduces blood viscosity. The aim of the present study was to assess the effect of increased fluid intake on blood viscosity. Men (n 67) and postmenopausal women (n 27) with one or more risk factors for CVD who reported intake of < or =0.5 litres water daily were randomised to a control group (n 31), an intervention group (n 32) that increased their daily water intake by 1 litre/d and an intervention group (n 31) that ingested 1 litre blueberry juice/d. All were encouraged to continue their usual diet and lifestyle. Whole-blood viscosity and blood and urine chemistries were measured by standard techniques after 2 and 4 weeks. Urine volume increased (by a median of 872 and 725 ml in the water and blueberry juice groups, respectively, v. 10 ml in the control group; P< or =0.002), confirming the subjects' adherence to the protocol. Urine osmolality and urinary levels of Na, K and creatinine decreased in the water and blueberry juice groups v. the controls (P<0.05). No change was seen in whole-blood viscosity or in levels of fibrinogen, total protein, lipids, glucose, insulin, C-peptide or other chemistry and haematology variables. In conclusion, a postulated protective effect of increased water or fluid intake is not explained by a change in blood viscosity and increased fluid intake does not influence CVD risk factors in the short term.