3

The Case against Eating Less Salt

The vast majority of Americans can take comfort in knowing their chosen dietary sodium intake is not a health hazard.

—Michael H. Alderman, MD,1 Albert Einstein College of Medicine

Most health authorities have concluded that as people increase their sodium intake their risk of heart attacks and strokes also increases. That’s what has led governments and health organizations around the world to urge people to eat—and companies to make—less-salty foods. The recommendations vary, depending on the organization, but they generally advise that people limit their sodium intake to less than 2,000 or 2,300 mg per day. In contrast, Americans’ current intake is around 3,400 mg. To most hypertension and cardiovascular disease experts, the case against salt was closed decades ago.

But there’s an alternative universe of thinking. As long ago as 1980, a minority of researchers opposed reducing sodium consumption. Citing conflicting evidence on the effects on health of moderate sodium reductions, John D. Swales, a medical professor at England’s University of Leicester, warned against “such massive public health measures as reducing the sodium content of food.”2 (It was later revealed that Swales was secretly working with the British salt industry.)3 A few years later, an international group of researchers wrote: “There is no scientific evidence” that reducing sodium intake would benefit the general public. The group further claimed that population-wide reductions were “unjustified and irresponsible.”4 That was more than a decade after the White House Conference on Food, Nutrition, and Health recommended that sodium consumption be reduced.5 Still, research then was not as fully developed as now, and delving more deeply into sodium’s effect on blood pressure and health made sense.

Similar criticisms have continued to this day, with some newer studies suggesting that current sodium intakes are optimal and that cutting back would be useless or even harmful. But even critics of reducing sodium across the population agree that people with enormous intakes of sodium, such as more than 5,000 mg per day, should cut back.

Andrew Mente, a nutritional epidemiologist and sodium-reduction critic at McMaster University in Hamilton, Ontario, said, “The bottom line is that there is not a shred of evidence whatsoever that low sodium, 2,500 mg per day or lower, is better than average sodium, around 3,500 mg per day, in reducing cardiovascular events or mortality.”6

In 2016, David A. McCarron and Michael H. Alderman, both long-time opponents of lowering sodium intakes, wrote in the Journal of the American Medical Association: “The general population’s greatest risk is at intakes below 2800 to 3000 mg/d. . . . The proposed FDA target of 2300 mg/d is significantly below that lower limit and thus unsafe.”7

I asked Alderman, now an emeritus professor at the Albert Einstein College of Medicine, for his current views on the salt debate. When he was starting to do research in the early 1970s, he told me, “I, like everybody else in the world, knew that a low-salt diet would lower blood pressure [and] was really a good thing to do.”

But [now] I think the issue . . . is really silly. There is not any evidence that reducing sodium to less than 2,300 milligrams per day is a benefit. . . . Why should we ask millions of people to change their diet when there is no evidence? I mean, it seems to me you need strong evidence to do something. Our evidence isn’t perfect, but I’m not asking anybody to change anything.8

Alderman also told me he’s convinced that people can’t change their sodium-intake habits, even if it were beneficial to consume less sodium.

Scientists have long debated the health effects of salt at countless conferences, at government advisory committee meetings, and in the pages of scientific journals. But in recent years, journalists at prominent news outlets—with their voices augmented by social media and bloggers—have broadcast those debates to the public in the form of “man bites dog” stories. Those articles have fueled confusion, leaving people buffeted by seemingly endless arguments between two camps of credentialed scientists:

And those articles have “legs.” Jeremiah Stamler, the Northwestern University epidemiologist, complained to the publisher and top editors of the New York Times about a Times article that was syndicated nationally and headlined “Hypertension Research Challenges Role of Salt.” In 1992, long before the internet and social media turbo-charged the art of propagandizing, Stamler observed that whenever an article questioning salt’s harmfulness was published,

within a short time it [was] sent, under the aegis and at the expense of private commercial and trade associations, all over the country, to doctors, researchers, nutritionists, etc. Press conferences are held, exhibits that misrepresent research findings are prepared and circulated. In short, efforts are made to make the health questions take a back seat, in favor of commercial interests.18

Even some medical journals publish papers that are more attention getting than reliable. Maybe that attracts readers, “clicks” on the web, and more advertisers, but it certainly does not serve the public interest.

What had seemed to be well-settled science became controversial, at least in the United States, though not much elsewhere in the world. It happened with climate change, it happened with cigarettes, it happened with lead. Is it now happening with salt? Or have the sodium skeptics truly proven that Americans are eating an optimal level of sodium, that lower-sodium diets would be harmful, and that government should not press companies to lower sodium levels?

Way back in 1989, a committee of the National Academy of Sciences, in its report titled Diet and Health: Implications for Reducing Chronic Disease Risk, recognized that diets high in sodium and low in potassium increase the risk of hypertension. But the committee also stated:

By far the greatest difference of opinion, and the most strongly held opinions, relate to the desirability of recommending to the general public that dietary sodium intake should be restricted. . . . There is little likelihood that these controversies will be entirely resolved in the foreseeable future.19

How right they were! Let’s now examine some of the pivotal studies that defenders of salt have cited when they proclaim that eating a lower-sodium diet would be worthless and even dangerous.

People with Normal Blood Pressure Need Not Consume Less Sodium

A linchpin of the plea by public health experts to lower sodium throughout the population is that doing so would lower blood pressure and prevent heart attacks and strokes. But what if lowering sodium had no effect on blood pressure in most people?

In 2019, Niels Graudal of the Copenhagen University Hospital in Denmark and several colleagues conducted a large meta-analysis of controlled clinical trials on sodium and blood pressure.20 Coincidentally, it included 133 studies, the same number as were included in a 2020 meta-analysis that I described in chapter 2.21 Like the subsequent study, this one found that a major decrease in sodium intake decreased systolic blood pressure only slightly (–1.46 mm Hg) in the majority of people whose blood pressure was normal, under 132 mm Hg. (The change in people with higher blood pressures was far higher: 7.7 mg Hg.) The authors used the smallness of the increase to argue that sodium reduction “should probably not be a target for the general population but only for hypertensives with a high sodium intake.”

The conclusion that most people need not reduce their sodium intake was seriously misguided. First, as I also noted in chapter 2, even small reductions in blood pressure averaged over the millions of people with normal or high-normal blood pressure would prevent a modest number of heart attacks and strokes in the coming decades. Second, people who have hypertension cannot lower their sodium intake significantly over the long term unless sodium is decreased in the overall food supply. Third, reducing sodium would recalibrate Americans’ taste buds, starting in childhood, and help reduce the taste for salt and the risk or severity of hypertension. Finally, the authors ignored the likelihood that salt may well be harmful by mechanisms other than boosting blood pressure and that elevated blood pressure causes problems other than cardiovascular disease.

The “Earth-Shattering” Institute of Medicine Report

The heart of sodium skeptics’ argument, though, does not rely on the effects of sodium on blood pressure, but rather on disease. In 2013, the Institute of Medicine (IOM; now the National Academy of Medicine or NAM) published a major, attention-getting report on how low-sodium diets might affect health.22 The journalist who wrote about the report in the New York Times—under the sensationalized headline “No Benefit Seen in Sharp Limits on Salt in Diet”—said it “undercuts years of public health warnings” and stated that Alderman called the report’s findings “earth-shattering.”23

The IOM committee was established to consider whether diets in the range of 1,500 to 2,300 mg of sodium per day affected health outcomes, such as strokes, instead of just risk factors such as blood pressure. The committee supported lowering sodium consumption from today’s high levels to 2,300 mg per day, yielding a substantial benefit and causing no harm.24 But it also concluded that the evidence was inconsistent and insufficient because of the paucity of evidence on whether intakes below 2,300 mg of sodium were beneficial or risky to the general population. Recall that the government’s then-current “Dietary Guidelines for Americans 2010” had recommended that older people, African Americans, and people with pre-hypertension and hypertension should shoot for 1,500 mg per day, something the makers of salt and salty foods did not appreciate.

Importantly, the committee did not evaluate the voluminous research showing that low-sodium diets, including ones with well under 2,300 mg per day, reduce blood pressure and that lower blood pressure reduces the risk of cardiovascular disease. According to the committee, that was not within the scope of the charge it was given by the government sponsors.

But here is the “earth-shattering” part of the evaluation, which led to huge publicity: the IOM committee stated that lowering sodium intake to 1,840 mg per day “may lead to greater risk of adverse events” in patients with heart failure. The word “risk” exerted its magnetic pull on journalists who discussed the IOM assessment.

The alleged risk was based largely on six studies done in Italy, where many patients with congestive heart failure (CHF) who were put on a low-sodium diet had died.25 It was not widely publicized that the doctors followed an unwise therapeutic regimen that was rarely used (and probably not at all in the United States). The IOM qualified its statement by noting that the patients whose sodium intakes were restricted had severe heart failure, had been subject to “aggressive” treatment with high doses of a powerful diuretic, and were on fluid-restricted diets. The American Heart Association (AHA) dismissed the Italian research entirely—and the committee’s reference to it—by emphasizing that “experience in such a sick and highly medicated group has no relevance for the general population or even for most patients with heart failure.”26

Putting a nail in the Italian research’s coffin, the journal Heart retracted (that is, it disavowed) a meta-analysis it had previously published of those six studies. The journal’s ethics committee stated that two of the studies contained duplicate data as well as raw data that could not be substantiated because, the researchers claimed, it had been “lost as a result of computer failure.”27 That’s either a good example of “the dog ate my homework” or an inexcusable failure to properly store data, which is further evidence that the study was poorly conducted. Before the meta-analysis had been retracted, the pro-salt health journalist Gary Taubes published “Salt We Misjudged You,” a prominently placed opinion piece in the New York Times that gave national publicity to the Italian research. Taubes wrote: “Italian researchers began publishing the results from a series of clinical trials, all of which reported that, among patients with heart failure, reducing salt consumption increased the risk of death.”28 Actually, it wasn’t the lack of salt that killed the patients; it was the doctors’ risky therapeutic regimen.

The IOM committee also found “some evidence suggesting risk” from consuming less than 2,300 mg of sodium per day for people with diabetes, chronic kidney disease, or preexisting cardiovascular disease. The evidence of risk to those various groups of patients was skimpy, but in any case patients with serious illnesses are ordinarily under their doctors’ care and are very different from healthy consumers, at whom public-health dietary advice is directed.

Dariush Mozaffarian, the dean of the Friedman School of Nutrition Science and Policy at Tufts University, later sharply criticized the 2013 IOM committee, saying that it

was not tasked with reviewing all available evidence nor with setting a target level. Rather, they were instructed to limit their focus to studies of clinical endpoints, and only to studies published from 2003 to 2012 . . . and only to the question of comparing a target level of 2,300 to 1,500 mg/day. Their task, in other words, was not to determine the best evidence base for a dietary target, but to evaluate one type of the evidence and over a specified period and only for the question of lowering the target from 2,300 to 1,500 mg/day.29

Controversy aside, the 2013 IOM report endorsed long-standing advice to cut sodium to 2,300 mg per day, but it did not support going below that amount. Truth be told, debating whether people should consume 1,500 or 2,300 mg is currently a bit academic considering that the average American consumes so much more—3,400 mg. It’s going to be a long, long time before the country gets down to an average of 2,300, let alone 1,500. But you, of course, don’t have to wait to consume less salt. I’ll have tips for doing that in chapter 11.

Similar to the IOM committee, in 2017 a joint committee of the World Heart Federation, European Society of Hypertension, and European Public Health Association pointed to the absence of controlled trials on the health effects of diets in the range of 2,300 mg per day. They advised remedial actions only when a population’s average sodium consumption exceeds 5,000 mg per day.30 That would give the green light to almost everyone in almost every nation to eat just about all the salt they want—with disastrous consequences for public health. In light of the mountain of research on sodium and disease, their advice should be ignored!

Observational Studies Suggest That Low-Sodium Diets Promote Cardiovascular Disease

Perhaps the most widely publicized evidence that low-sodium diets could be harmful was based on studies of sodium intakes of large groups of people over a number of years. In such observational studies the participants are not asked to change their diets, they are just observed. In contrast, people in trials are put on diets (or urged to adhere to certain diets) with different amounts of sodium and then followed for months or years to identify any differences in blood pressure or rates of disease or deaths. Observational studies probably (and unfortunately) have muddied the waters and confused the salt debate rather than clarified it. Because this issue is at the heart of the debate over how much sodium we should consume, I am going to delve deeply into it.

Observational studies have been highly controversial: although some found a lower risk of cardiovascular disease at lower sodium intakes (see chapter 2), others linked a higher risk of cardiovascular disease to sodium intakes at both higher- and lower-than-typical intakes.

Some major observational studies are based on dietary intakes measured by the federal government’s National Health and Nutrition Examination Survey (NHANES). Researchers mine those data for all sorts of relationships between diet and disease. Hillel W. Cohen and Michael H. Alderman of the Albert Einstein College of Medicine and their co-authors have published several papers based on NHANES surveys conducted in different years. They consistently found that the lowest sodium intakes were associated with a higher risk of deaths from cardiovascular disease. For instance, a 2006 article concluded that consuming less than 2,300 mg of sodium per day was associated with a 37 percent greater risk of cardiovascular disease and a 28 percent greater risk of dying from any cause compared to people who consumed more than 2,300 mg per day.31

Stated plainly, those investigators found that the people consuming the least sodium and the most sodium had a greater risk of cardiovascular disease than those consuming middling levels. The authors acknowledged certain inherent weaknesses in their and other observational studies, but the unexpected results, published in major medical journals, lent credence—and publicity—to the notion that people need not lower their sodium intake. To overcome those weaknesses, Alderman and others have urged that controlled trials be done in which large numbers of people would be asked to eat, or be provided with, diets with different levels of sodium. Then they would be followed for years to determine the rates of heart disease, strokes, and overall deaths. I write more about that later in this chapter.

Feng J. He and her fellow researchers in London blasted the Cohen-Alderman study. They decried that “the method used to assess salt intake (one 24-hour dietary recall at the beginning) is notoriously unreliable, particularly because no account is taken of discretionary [table] salt.” And, because Cohen’s group had conducted earlier studies using the same flawed methods, they added, “It is quite extraordinary that Cohen et al. choose to ignore the scientific criticisms that followed their [previous] article.”32

Alderman is not totally doctrinaire about the harmfulness of lowering sodium. He has acknowledged that salt restriction, though “relatively weak and costly,” does lower the blood pressure of some patients.33 And he agrees that it would be sensible for people with enormous intakes or who have hypertension to reduce their intake. I asked him if he ever has doubts about his position that current salt intake is generally fine, especially when health organizations with the stature of the World Health Organization (WHO) and Centers for Disease Control and Prevention (CDC) strongly favor public health measures to lower their sodium intake population-wide. He said “of course.”34

But Alderman questions whether it is even possible for people to consume less salt—because, he says, decades’ worth of advice to the public to cut the salt has had no effect. That argument is weak at best. Official policy is to cut the salt, but the US government has never mounted a well-funded, persuasive, persistent education campaign. Relying on a standard low-budget and perfunctory education program to lower sodium is like using scissors to cut the greens on a golf course. Living in a world of salty packaged and restaurant foods makes it very challenging for average consumers, who have many more immediate worries on their minds, to opt for a lower-salt diet.

Critics charge that many observational studies do not just have weaknesses, but are so flawed as to be misleading. Finnish researchers said about earlier (1978) research, “Rather than shed new light on sodium intake and mortality, Alderman and colleagues’ report brings unnecessary confusion into the discussion on the relation between dietary sodium and mortality.”35 They pointed out that the people who had supposedly consumed low levels of sodium had “a calorie intake that should have resulted in death from starvation.” People in the lowest one-fourth of sodium consumption reported that they consumed only about half the recommended calorie intake. Clearly, those people were underestimating how much food (and, hence, sodium) they had eaten.

Fatal and PURE Flaws

In a systematic critique of the research indicating that low sodium intakes were harmful—with most of that assessment coming from observational studies—CDC researchers, including CDC’s then-director Tom Frieden, emphasized the studies’ often-poor estimates of sodium intake.36 They noted that basing sodium intakes on participants’ recollections of what they ate on just one day at the beginning of a long study is unreliable and might well lead to inaccurate results. People’s diets vary radically from one day to another (just think of your own diet). Moreover, people tend to under-report the soups, restaurant meals, and other unhealthy foods they ate and over-report the broccoli, spinach, peaches, and other healthy foods. Even Alderman and two colleagues acknowledged that measuring sodium intakes just once in a multiyear study, as is done in much observational research, might not be adequate.37

The CDC scientists also noticed that a disproportionate number of participants consuming a low-sodium diet had diabetes, hypertension, heart disease, or other chronic illness when the studies started. But sick people eat less food and therefore less sodium, and some of those people likely had consumed less salt to help treat their illness. It makes no sense to assume that low sodium intake caused them to be sick. That kind of “reverse causality,” also called “reverse causation,” confuses cause and effect and is an inherent defect of a great deal of observational research.

To get beyond individual observational studies, Graudal, Alderman, and two other researchers conducted a meta-analysis based on some two dozen observational studies on sodium consumption and the risk of cardiovascular disease.38 They found that the lowest intakes of sodium (under about 2,600 mg per day) and highest (above 4,900 mg per day) were associated with a higher risk of cardiovascular disease than intakes between those levels. That is, they observed a J- or U-shaped relationship between sodium intake and the incidence of disease, as depicted in figure 3.1.

Figure 3.1

Both high- and low-sodium intakes appear to be harmful. Several observational studies found a J- or U-shaped relationship between sodium intake (along the horizontal axis) and the risk of cardiovascular disease. In this hypothetical example, the lowest risk of disease was with a daily sodium intake of 4,500 mg. Source: Illustration by J. Bach, CSPI.

Findings that are so contrary to the larger body of research on salt and disease must be scrutinized carefully. In fact, meta-analyses, which are sometimes considered especially reliable because of their increased statistical power, may distract from weaknesses in the underlying individual studies and, hence, in the meta-analysis itself. When the indefatigable British researchers Feng J. He and Graham A. MacGregor put the Graudal meta-analysis under their microscope, they stated that its “conclusion is invalid because of the severe methodological flaws of the studies [it] included.”39 They pointed especially to inaccurate measurements of sodium consumption (“it varies hugely day to day”), as well as the likelihood of reverse causality. More recently He and her colleagues wrote: “These J- or U-shaped findings should not have been used to challenge the current public health policies due to their severe methodological limitations.”40

By coincidence, the Graudal meta-analysis of observational studies was published in 2014 just after the American Heart Association (AHA) had released a detailed “science advisory” that explained why many such studies are unreliable.41 The advisory emphasized, like He and MacGregor, erroneous measurements of sodium intakes and reverse causality. The AHA advisory concluded, “It remains appropriate to base [sodium] guidelines on the robust body of evidence linking [sodium] with elevated blood pressure and the few existing general population trials of the effects of [sodium] reduction on [cardiovascular disease].”

The giant among observational studies is the Prospective Urban Rural Epidemiology (PURE) study—actually a series of studies begun in 2001 and led by Salim Yusuf of McMaster University Medical School in Hamilton, Ontario. PURE is frequently portrayed as being exceptionally powerful and reliable because of its enormous size. For example, for a 2014 study, the researchers collected urine samples from 101,945 persons in 17 countries to determine sodium intake, and then followed individuals for three-and-a-half years.42 They found that consuming between 3,000 and 6,000 mg of sodium per day was associated with a lower risk of death and fewer heart attacks and strokes than consuming a higher or lower intake.

In a subsequent and even larger analysis involving 133,000 individuals, PURE researchers found the same J-shaped relationship between sodium consumption and the risk of cardiovascular disease.43 (Other PURE studies found that saturated fat was not harmful and perhaps even beneficial, and that polyunsaturated oils and vegetables were not beneficial; both findings are contrary to most medical research.)44

Martin O’Donnell, a colleague of Yusuf, told one journalist that PURE undermines much of the advice that health officials have been telling the public for years:

This study . . . questions the appropriateness of current guidelines that recommend low sodium intake in the entire population. An approach that recommends salt in moderation, particularly focused on those with hypertension, appears more in line with current evidence.45

Graudal, who was not an author but also opposes lowering sodium intakes below about 3,000 mg per day, said flatly that PURE “is based on genuine scientific data” and “confirms that low sodium intake is an independent risk factor for increased mortality.”46 In other words, he implied, everything you’ve heard about the importance of cutting sodium well below the current 3,400-mg average diet is wrong. And journalists, impressed by PURE’s size, quickly pounded out prominent and favorable stories that spread the message. Unfortunately, such media coverage amplifies and gives credence to unreliable studies, and that can lead consumers to think that cutting back on salt is dangerous.

PURE’s conclusions were immediately challenged. Francesco Cappuccio, professor of Cardiovascular Medicine & Epidemiology at the Warwick Medical School in England and head of the WHO’s Collaborating Center for Nutrition, slammed both the PURE paper and the Lancet medical journal for publishing it. He told the Independent newspaper, “It is with disbelief that we should read such bad science published in The Lancet. . . . The flaws that were extensively noted in [the researchers’] previous accounts are maintained and criticisms ignored.”47

Nancy Cook, the biostatistician at Brigham and Women’s Hospital and Harvard Medical School, took on the issue of study size:

The fact that the PURE study is the largest to date should not influence its interpretation. A large study size does not eliminate bias resulting from selection, reverse causation, or confounding [i.e., interference by a third variable] but could lead to spurious results.48

The observations of the heart association’s science advisory are particularly applicable to PURE: “It is difficult to conduct rigorous, high-quality investigations of the relationship between [sodium] intake and [cardiovascular disease].” The advisory included a further comment: “For the foreseeable future, the high-quality body of evidence linking [sodium] intake to [blood pressure] should remain the basis for setting recommended levels of [sodium] intake.”49

The PURE authors themselves acknowledged some “limitations” in their methodology, and many critics heartily agreed. The “limitations,” though, were disqualifying flaws. One of the biggest was the reliability of participants’ sodium intakes. I touched on the problem earlier when discussing the Graudal study, and it’s worth exploring here in more detail.

Instead of obtaining multiple 24-hour urine samples to estimate usual sodium intakes, PURE used just one “spot” urine sample—peeing into a container once—at the beginning of the studies. That method saves money, but it does not accurately reflect a person’s average sodium intake both at the beginning of a study and over time. (Intakes based on what participants said they ate on the previous day or based on which foods they reported consuming over the course of months—Food Frequency Questionnaires—are similarly flawed.) To correct for the measurement problem, the PURE investigators used a formula—the Kawasaki formula—to try to convert the sodium content of the spot urine samples to what would have been excreted over 24 hours. But the Kawasaki formula has been dubbed a “poor performer” because it overestimates sodium at lower levels and underestimates it at higher levels.50 Statistical alchemy simply cannot turn bad data into good data.

In addition, two reports demonstrated the perils of using spot urines and the Kawasaki formula. In one, Dutch researchers took advantage of their previous project in which people provided several 24-hour urine samples over 15 years.51 That study found that the average sodium content of multiple urine samples was often markedly different from the initial sample. When they used only the initial 24-hour urine sample to correlate sodium intake with cardiovascular risk—note that even a single 24-hour sample is more reliable than the spot urines in PURE—the notorious J-shaped relationship emerged. But they then included urine samples obtained after one year and five years to get a truer estimate of a subject’s typical sodium intake. Like magic, the J-shaped curve indicating a greater risk at the low end of sodium consumption vanished. Instead, the relationship between sodium and cardiovascular disease became the expected linear one, with the lowest sodium intakes being associated with the lowest risk of disease. The Dutch scientists warned against relying on a “wobbly parameter” (a single urine sample) and took a direct poke at PURE’s basic methodology: “Future [observational] studies should therefore focus on accurate assessment of sodium intake rather than the inclusion of many subjects.”52

In a second rebuttal, a team of British, Canadian, and American experts also disputed the reliability of PURE. They did a reanalysis based on the two TOHP studies that I discussed in chapter 2, which found that the participants with a lower sodium intake had a lower mortality rate. Similar to the Dutch study, when the team reanalyzed their data using the Kawasaki formula to estimate sodium intake, the J-shaped curve appeared. That finding, like the Dutch one, demonstrated that the curve was an artifact of PURE’s methodology and not a true indication that low salt intakes are harmful. The researchers’ stinging conclusion was that “paradoxical results from methodologically flawed studies should not be used to derail critical public health policy, nor divert action.”53

Most damning, the TOHP researchers investigated whether participants in their control group who consumed the least sodium, under 2,300 mg per day, had higher rates of cardiovascular disease than people who consumed middling levels.54 In fact, they found the opposite: people who consumed less than 2,300 mg had a 32 percent lower risk than people who consumed 3,600 to 4,800 mg. That finding was shy of statistical significance, but that might have been due to the small number of participants who consumed so little sodium. As the researchers concluded, “estimates from spot urine are unreliable, not reproducible and systematically biased.”

Cook highlighted yet another limitation of PURE and other observational studies: “residual confounding.” That is the term researchers use for unaccounted-for factors that can distort the apparent relationship between an exposure and a health outcome. Because they are unknown, researchers cannot compensate for them by making statistical adjustments. Residual confounding is an especially important problem, Cook said, in a study that includes “people from a host of countries, ranging from low to high income, with very different background health status, nutritional standards, and health care systems.” Cook further explained:

[The PURE researchers] only control for a few variables and don’t capture the heterogeneity in the data. For example, some of the participants could even be malnourished or have other uncontrolled health conditions including infectious diseases, and that could easily account for the effects seen.55

Yusuf, the lead PURE researcher, has acknowledged the problem: “Even the best work has limitations, including ours. So, I worry how much of our work is potentially confounded. I truly worry.”56 In a presentation to the National Academy of Medicine, he said, “That’s our data. Would I like better data, yes.”57

Still, the PURE researchers continue to campaign against sodium reductions. Their persistence has driven some leading hypertension experts to frustration. In a detailed, scholarly rebuttal to the sodium skeptics, Norm R. C. Campbell, a professor at the University of Calgary’s O’Brien Institute of Public Health and Libin Cardiovascular Institute of Alberta, charged:

Several dissenting scientists have conducted low-quality research, taken research out of context, made factual errors or misinterpreted results, altered scientific formulae/protocols in a fashion that makes their controversial research appear more robust, and used low-quality evidence to trump higher quality.58

More colloquially, a quip favored by software engineers comes to mind. When you combine data from rich and poor countries and healthy and sick people into one big study, you end up with “garbage in, garbage out.”

No matter how big the studies and how hard the researchers worked on them, the PURE studies (and others conducted like them) are red herrings, especially in the context of the vast body of reliable research on sodium and health.59 The AHA told consumers bluntly, “The findings in this [PURE] study are not valid, and you shouldn’t use it to inform yourself about how you’re going to eat.”60

The “Set-Point” Theory of Sodium Consumption

In 2013, based on their reviews of sodium consumption in 45 countries, several widely published researchers have asserted that the “‘normal’ range of human sodium intake [is] defined by physiology and biological needs and not by the food supply.”61 Those researchers—McCarron, a consultant who previously headed the nephrology division at Oregon Health & Science University, Niels A. Graudal in Copenhagen, and others—found that average sodium consumption in almost all countries ranged from 2,600 to 4,800 mg per day. They also noted that in the Trials of Hypertension Prevention (TOHP) II trial (see chapter 2) participants were vigorously counseled to consume about 1,800 mg of sodium per day, but they actually consumed closer to 2,600 mg.

Based on such observations, the researchers suggested that humans have a safe and natural “set point” for sodium consumption that is determined by hormones that regulate the excretion of water and sodium and not by what people try to consume or what the food industry is marketing. They also raised the specter of harm if consumption dipped below 2,300 mg per day, saying, “to attempt to use public policy to abrogate human physiology would be futile and possibly harmful to human health.”62

The true situation is more complicated. It may sound reasonable to say that people around the world are consuming the optimum level of sodium, but it is misleading.63 For starters, McCarron and his colleagues excluded from their study populations that consume very salty diets, as well as groups that consume very little sodium, such as hunter-gatherer tribes with intakes amounting to just a few hundred milligrams per day. Also, because salt is such an integral ingredient of packaged and restaurant foods throughout most of the world, consuming at least 3,000 mg of sodium per day is not a conscious choice or a physiological requirement, but an almost inevitable consequence of our salty food environment.

What really puts the lie to the set-point theory of sodium consumption is that people can consume less sodium. In Finland, average sodium consumption was cut by one-third. In a controlled experiment in two small Portuguese communities, people in the intervention community reduced their sodium intake by 43 percent after two years.64 Average consumption in the United Kingdom was cut by 10 to 15 percent over a decade. The participants in the TOHP and TONE trials reduced their intake by one-third, with those in TOHP I consuming 2,300 mg per day. Admittedly, though, no entire country has cut sodium consumption all the way down to 2,300 mg per day or less. That’s impossible to do when packaged and prepared foods are suffused with salt and other sources of sodium.

Yet More Criticisms of Low-Sodium Diets

Might choosing a low-sodium diet somehow lead to nutrient deficiencies? Two University of Washington nutrition researchers were curious to see how practical it would be for Americans to reduce sodium intakes substantially and still consume adequate amounts of protein, vitamins, and other nutrients. Matthieu Maillot and Adam Drewnowski used a mathematical model to estimate the dietary changes that would be needed to consume a diet with 2,300 or 1,500 mg of sodium per day accompanied by 100 percent of the recommended intakes of two-dozen other nutrients. They concluded that eating a nutritionally adequate diet with as little as 2,300 mg of sodium was feasible for most men and women. Yet to get down to between 1,500 and 2,000 mg, they said, would require “wrenching” changes in food choices and the American food supply. To get to 1,500 mg or less would necessitate totally omitting meats and grains. Maillot and Drewnowski stated: “No combination of food categories satisfied the model requirements of a nutrient-adequate food pattern. . . . In other words, at this low level of sodium, the requirements for multiple other nutrients could not be met.”65

But the situation is not as dire as the University of Washington researchers suggest. The 2010 Dietary Guidelines Advisory Committee described food patterns with 2,300 mg of sodium and 100 percent of the recommended intakes for almost all nutrients (exceptions were vitamins D and E).66 To adhere to those dietary patterns, however, most people indeed would have to eat more natural, whole foods and less (and less-salty) packaged and restaurant foods—changes that would be salubrious for many reasons. Getting down to 1,500 mg a day would certainly require greater changes in food composition and choices, but for now let’s be satisfied getting down to 2,300 mg in the next decade or so.

In another line of attack, some researchers have suggested mechanisms by which low sodium intakes could be harmful. One possibility, they contend, is that lowering sodium consumption could upset hormonal balances. They focus on the body’s renin-angiotensin-aldosterone system (RAAS), a group of critically important hormones that regulate sodium and fluid balances in the body. Alderman wrote that at sodium intakes under 2,500 mg per day, plasma renin activity increases and mortality increases.67

Indeed, major, sudden reductions in sodium consumption sharply increase levels of renin and aldosterone, which could be harmful. But over the longer term and with more modest (and realistic) reductions in sodium, renin and aldosterone levels stay about the same.68

Cappuccio, the head of the WHO’s Collaborating Center for Nutrition, told National Public Radio that elevated renin-angiotensin activity is the body’s normal physiologic response to decreased sodium, and is not worrisome.69 The Yanomami Indians, as a result of their extremely low sodium intakes, have renin levels 10 times higher than Americans’ levels without any apparent problem.70 It is likely that those are historically normal levels, and that our low levels of renin and aldosterone are aberrant. Also, taking diuretics to treat hypertension stimulates the renin-angiotensin system, but diuretics are known to reduce cardiovascular mortality.71 Finally, in 2010 the IOM said, “in contrast to the well-accepted benefits of blood pressure reduction, the clinical relevance of modest rises in plasma renin activity as a result of sodium reduction is uncertain.”72 So let’s not worry that gradually declining sodium intakes might lead to dangerously high levels of those hormones.

The Dangers of Waiting for the Perfect Trial

Sodium skeptics have opposed lowering sodium consumption by the general population until definitive research has shown that reductions would be safe and effective in reducing disease rates. Alderman urges that before health officials take actions to lower sodium, “all researchers should press for well-designed, rigorous, and robust [randomized controlled trials, or RCTs] to determine the health consequences of universal salt restriction.”73 Alderman and several others have been saying that for at least 30 years.

In theory, definitive RCTs to determine whether sodium intakes under 1,500, 2,300, or 3,000 mg per day decrease or increase the risk of disease would be the ultimate test of the “sodium hypothesis.” The ideal trial would enlist a large number of healthy volunteers, split them into two groups, and for many years give them all their meals and snacks, which would be identical except for differences in sodium content. Then the researchers would compare the numbers of strokes, heart attacks, and other health problems that people in each group suffered.

Partly because of the huge cost of that kind of trial, almost two decades ago two prominent experts predicted flatly, “it will never be done.”74 Many researchers, who almost reflexively support doing more research, are not supportive here. According to Lawrence J. Appel, a professor at the Johns Hopkins School of Medicine and one of the lead researchers in the DASH trials, “it would not be worth the considerable time and expense because of the overwhelming evidence for salt’s adverse effects on blood pressure.”75 I suspect that because the National Institutes of Health and other agencies around the world have not provided funding, they have apparently come to the same conclusion.

MacGregor emphasizes that nutrition research is often not like drug research where randomized trials are generally feasible. He likened the situation to that of tobacco, where we have no trials showing the benefits of smoking cessation. Instead, he said, “We need to rely on all the other types of evidence, and for salt we are fortunate to have over 10 different types of evidence, all of which indicates that salt is important in increasing blood pressure.”76

Pasquale Strazzullo, a professor at the University of Naples Medical School in Italy, echoed that sentiment, asking, “Should we refrain from this life-saving measure and let people die of hypertension and its cardiovascular complications while waiting for the ‘mother of all trials’?”77 Health officials should not be paralyzed by the paucity of trials.

Despite the obstacles to conducting a definitive trial, in 2018, eight well-known researchers, including both advocates and opponents of reducing sodium intakes—such as Paul Whelton of the Tulane University School of Public Health on the pro side and McCarron and Alderman representing the cons—explored options for conducting an RCT with prisoners.78 They wanted to work with prisons to provide some inmates a diet with the usual amount of sodium and to give other prisoners, in the same or different prisons, the exact same diet but with less sodium. Because the Federal Bureau of Prisons does not allow research other than that which advances knowledge about corrections, such a study would probably have to be conducted in state or private prisons.79

Some public health experts, however, were leery. They contended that an RCT involving prisoners was (a) unnecessary (because of the strong and consistent animal, clinical, and epidemiologic research, the several existing trials, and the experience in the United Kingdom and Finland showing the benefits of lowering sodium intakes); (b) too expensive (and potentially siphoning funding from higher-priority research); (c) possibly unethical; and (d) almost certainly inconclusive, leading to calls for additional costly and lengthy trials.

Sonia Y. Angell, then a deputy commissioner of health in New York City and now the director of the California Department of Public Health, pointed out that “prison conditions are unique,” and prisoners not only have higher rates of mental illness, cardiovascular disease, and other problems, but they also experience stress levels that the general population does not endure. Angell said, “Modifying a single nutrient in a diet of a population in prison won’t produce a study with answers relevant to our population at large.”80

Ethical considerations raised additional concerns: Would the inmates truly be giving their informed consent to participate in the study? Should some of the participants be forced to eat a low-sodium diet that they would rather not eat? Should others be required to eat a standard, unhealthy, high-sodium diet?

Finally, and importantly, postponing public health action pending the results from an RCT could undermine public health because, according to Appel, planning, conducting, analyzing, and publishing a study might delay policies to lower sodium in the food supply for 15 years.81 During that time companies and supportive politicians would likely argue that the government should not encourage consumers or require companies to reduce sodium until the results were available. (My organization, the Center for Science in the Public Interest, did not oppose the study, but pointed to the ethical, practical, and financial obstacles to conducting it. We also said that any such research should not be permitted to delay government action on salt.)

Ultimately, according to two members of the committee who spoke to me confidentially, the notion of a prison study fizzled when the advocates could not develop a design that would pass scientific and ethical muster, let alone be financially feasible. Even Alderman, who was on the committee and has called for trials for 30 years, stated in August 2019: “I don’t think it’s possible” to conduct an RCT.82 One insoluble problem he pointed to was the prisoners’ uncontrolled access to the commissary, and the customary selling and bartering of food they can buy there or receive from visitors. The demise of the prison study might have silenced calls for a gold-standard trial that had the potential to end the controversy. But it didn’t. When I talked to Alderman three months later, he had a new suggestion for an RCT, this time one in which people already consuming a low-sodium diet would be given either a salt tablet or a placebo and then monitored for heart attacks and strokes. Again, practical and ethical problems would certainly sink such a trial.

It is worth recognizing that public health measures are often taken in the absence of randomized controlled trials. Policy makers must rely instead on animal, epidemiology, clinical, and other evidence. Health officials have advised people to lose weight, stop smoking, avoid trans fat, and eat more fruits and vegetables, all without robust, controlled trials.

If asbestos researchers were to look at the battle over salt, they might quote the great New York Yankees catcher Yogi Berra, who purportedly said, “It’s déjà vu all over again.” The asbestos industry defended the safety of its product for decades after it was known to cause cancer (again, no randomized controlled trials). David Egilman, a clinical professor of family medicine at Brown University and then editor of the International Journal of Occupational and Environmental Health, described the asbestos industry’s battle plan this way: “They can throw a lot of things at the wall and hope something sticks with the jury. . . . It forces people like me or other scientists to try to clean up each thing that was thrown at the wall, one at a time. And by the end of the day, that could be confusing to a jury or judge.”83 Sounds like salt to me.

Summarizing the Science

Most public health officials recognize that lowering sodium intake helps prevent cardiovascular disease. But some scientists have conducted large observational studies—that is, not controlled trials—that appear to indicate that consuming less than about 4,000 mg of sodium per day increases cardiovascular disease. That finding is welcomed by the salt and food industries and publicized in the media, but the research has been roundly criticized, including by an authoritative 2019 committee of the National Academy of Medicine.

The scientists who oppose lowering sodium insist that new, long-term, controlled trials must be done before consumers reduce their sodium intake and governments adopt policies to help them do so. (Some of the scientists who oppose lowering sodium have received small amounts of industry funding and collaborated with the food and salt industries; I address those situations in chapters 5 and 6.) But during the several decades in which researchers have called for such trials, the preponderance of persuasive evidence showing how lower sodium intakes would be healthful, not harmful, has apparently convinced funders that such expensive projects are unnecessary.

Whether or not you followed every twist and turn of the research I’ve described so far, I’ll cut to the chase in chapter 4, where I summarize the evidence for consuming more sodium or less.