OBJECTIVE--To estimate the quantitative relation between blood pressure and sodium intake. DESIGN--Data were analysed from published reports of blood pressure and sodium intake for 24 different communities (47 000 people) throughout the world. MAIN OUTCOME MEASURE--Difference in blood pressure for a 100 mmol/24 h difference in sodium intake. Allowance was made for differences in blood pressure between economically developed and undeveloped communities to minimise overestimation of the association through confounding with other determinants of blood pressure. RESULTS--Blood pressure was higher on average in the developed communities, but the association with sodium intake was similar in both types of community. A difference in sodium intake of 100 mmol/24 h was associated with an average difference in systolic blood pressure that ranged from 5 mm Hg at age 15-19 years to 10 mm Hg at age 60-69. The differences in diastolic blood pressure were about half as great. The standard deviation of blood pressure increased with sodium intake implying that the association of blood pressure with sodium intake in individuals was related to the initial blood pressure--the higher the blood pressure the greater the expected reduction in blood pressure for the same reduction in sodium intake. For example, at age 60-69 the estimated systolic blood pressure reduction in response to a 100 mmol/24 h reduction in sodium intake was on average 10 mm Hg but varied from 6 mm Hg for those on the fifth blood pressure centile to 15 mm Hg for those on the 95th centile. CONCLUSIONS--The association of blood pressure with sodium intake is substantially larger than is generally appreciated and increases with age and initial blood pressure.