illustration of a salt shaker, shaking out salt

Salt and other wounds


That lovin' spoonful
Salt is poison
Salty history
Intersalt assault
Big study. Big problems?
Millimeters of megadeath
Finding common ground











salt shaker icon

The lesson of Intersalt was highly debatable







Some of these
cheeses pack
a hefty potion
of salt
















Scatter plot shows salt intake and blood pressure from 52 societies around the globe.
Now you see it, now you don't.
With data from four traditional societies
(four dots on left), Intersalt shows salt intake
and blood pressure rising hand in hand.
The bulk of the data show no relationship.
So what can you conclude?
Data from Intersalt

The last word?

By the early 1980s, the stage was set for a prolonged and expensive query into the role of dietary sodium. When it began in 1984, the Intersalt study was expected to provide the final word on salt and health.

With 10,000 participants from 52 communities around the world, it certainly had the potential to find a relationship between salt intake and blood pressure. More than anything, however, the study proved that salt's health impact remains elusive, debatable and sometimes contradictory.

As they expected, the Intersalt researchers found that higher salt intake was associated with higher blood pressure – individuals and societies with higher average intake also had higher pressure.

But when considering societies as groups, the relationship broke down. True, when four primitive societies with extremely low salt intake and extremely low blood pressure were included, the relationship existed. But when they were excluded, the other 48 study groups had no relationship at all.

Does this prove that sodium contributes to hypertension? Perhaps. But blood pressure and salt intake were only two differences between the four societies and the other 48. The four had less stress, less obesity, and ate far less fat and processed foods, and a raft more fiber. And on average, they died too soon to get much cardiovascular disease.

Mining salt data?
Part of the cheese aisle at a supermarket in America’s Dairyland.Given these considerations, critics charge that Intersalt actually backfired because the four societies with low salt and low blood pressure were too dissimilar to be compared with the other 48. "I cannot calibrate you to someone in a Brazilian tribe," says David Freedman, a statistician at the University of California at Berkeley. "This is apples and string beans, it's not comparing apples and oranges."

"Intersalt expected to find that high-salt centers had high blood pressure," Freedman continues, "and that's true in a crude sense. But the low-salt centers are very different. Blood pressure does not rise with sodium in the other 48 – in fact, as sodium increases, blood pressure decreases. The result in the 48 centers is exactly the opposite of what the Intersalt investigators expected to see," Freedman contends, and "the primary hypothesis of Intersalt was flatly contradicted."

salt shaker icon 2The Intersalt investigators then reanalyzed their data and announced that eating salt was correlated with a faster rise of blood pressure with age. Yet Freedman argues that this question was not an original goal of the research, and therefore is a case of "mining the data" to prove preconceptions rather than hypotheses.

To Freedman, the message is simple: "You should not be drawing dietary advice from this type of analyses." The Why Files tried to ask Intersalt investigator Jeremiah Stamler at Northwestern University about the study, but he didn't call back.

Curiously, the Intersalt study SEEB (see "Intersalt: An International ... " in the bibliography), was followed in the journal by a study of 7354 people in Scotland (see "Urinary Electrolyte Excretion... " in the bibliography)

What about getting sick and dying?


          back more
    The Why Files    
There are 1 2 3 4 5 6 7 pages in this feature.
Bibliography | Credits | Feedback | Search