Many public health agencies recommend universal restriction of dietary sodium to 100 mmol/L or less per day. This reflects the belief that because sodium restriction reduces blood pressure, it will therefore also reduce cardiovascular disease morbidity and mortality. Although large (100 mmol/L/24 h) reductions in sodium intake do produce a measurable decline in aggregate blood pressure, there is great heterogeneity in individual response, probably reflecting differing genetic, environmental, and behavioral characteristics. Moreover, sodium depletion has multiple other effects including activation of the renin-angiotensin system and the sympathetic nervous system, and increase in insulin resistance. Since the health effect will be the sum of these multiple good and bad effects of sodium reduction, outcome trials are needed to determine the benefit or harm of alteration in sodium intake. Unfortunately, no clinical trial has addressed the question of whether a lower sodium diet would improve or extend life. The best available data derives from six prospective cohort studies. In sum, the scant available observational data do not rule out the possibility of benefit for some and increased risk for others. Considerable experience--most recently the hormone replacement study--underscores the hazards of extrapolating clinical recommendations from observational data alone. In the absence of any evidence from randomized trials of morbidity and mortality outcomes, and in the face of inconsistent observational studies, a universal recommendation for sodium restriction is unwarranted and inconsistent with the principles of evidence-based medicine.