Several years ago, the massive Women’s Health Initiative examined the health outcomes of hormone replacement therapy (HRT) where post-menopausal women received supplemental estrogen. “Everyone” knew it was safe and it made logical sense: after menopause, women didn’t produce estrogen so, “of course,” replacing the hormone would make them healthier and live longer lives.

Only it didn’t.

Results of the trial showed clearly that many women died from the treatment. Surprise. The embarrassed NIH quickly shut off that portion of the trial and doctors were warned that what “everyone” knew was the right treatment regime was, in fact, endangering the lives of their patients.

The Salt Institute has cited the incident as a “learning experience” for public health policy-makers: that sometimes the most obvious and popular health nostrum turns a cropper so prudence dictates reserving population health interventions to those that have been tested in controlled trials. HRT never had been tested before it was rushed into practice. “People are dying. We can’t wait,” cried advocates.

Well it’s happened again. Same study. Different health outcome. A study published in the New England Medical Journal February 5 found increased incidence of breast cancer in HRT-treated women “suggesting a cause-and-effect relation between hormone treatment and breast cancer.”

Coming close on the heels of two studies done in Italy that found Coronary Heart Failure patients receiving salt-reduced diets (because for the past century “everyone knew” they work or at least cause no harm) suffered massively greater mortality than those on regular salt diets, the new HRT study drives home the point made repeatedly by the Salt Institute: we need a controlled trial of the health outcomes of low-salt diets. We should not ask the population to be the guinea pigs as we did the women in the Women’s Health Initiative.

Elsewhere we report the news that the European Union is cracking down on food health claims, approving only 9 of the first 43 evaluated and imposing stricter-than-FDA standards, usually based on "convincing" evidence from human clinical trials.

We applaud EFSA for its professed objective of preventing unproven health claims on food labels that would further confuse consumers in making informed dietary choices. To do the job right, EFSA has addressed one critical element: reliable, high quality data should be required to support approved health claims. There is a second crucial standard EFSA should also address: health claims should be based on disease outcomes, not intermediate variables, risk factors or biomarkers -- hard outcomes like disease-caused events or mortality.

On the first point, data quality, the key is to adopt a process like "evidence-based" medicine as defined by the Cochrane Collaboration: first set the standard for quality before looking at what the evidence might be; that's exactly the opposite of what the US Dietary Guidelines Advisory Committee is doing in its expert-based review (which it also calls "evidence-based").

For the second point, the key is to stop looking at "indicators" and "risk factors," recognizing that there are often multiple and conflicting bodily processes affected by any single nutrient. Not only is a carefully controlled trial needed to isolate those effects, but its is their combined totality that concerns us: does the intervention improve health, extending healthy lives, or just change one risk factor (ignoring others)?

With regard to salt, the implications are enormous. The Cochrane Review on health outcomes of salt-reduced diets finds insufficient evidence of improved health to justify a population advisory (as has the US Preventive Services Task Force). That much is clear, though ignored by diehard salt reductionist activists like WASH and its most vocal advocate, Dr. Larry Appel. It is the second question that will prove pivotal for EFSA if it hews consistently to the high standard it professes. For salt, the proper measure is cardiovascular health. Not blood pressure. Not insulin resistance. Not neurohormonal activity. All of these are risk factors. Some "improve" in some people when they reduce salt. Others impose additional risks when salt is reduced. It is the net effect -- the health outcome -- that is important.

In the US, FDA has wandered off into the health policy wilderness on this point. Its health claim (whatever one's view of its scientific rigor, the subject of another discussion) is for "hypertension," a risk factor, not "cardiovascular health" (or heart attack incidence, another "hard outcome"). FDA could just as well approve a health claim that diets with today's salt intake levels protect insulin resistance and warn consumers of the fact that salt-reduced diets endanger our ability to metabolize glucose properly, setting up low-salt dieters for diabetes and metabolic syndrome. For FDA to follow the professed EFSA approach, it should replace a "salt and hypertension" health claim with one for "salt and cardiovascular health" -- but then, of course, back to point one, quality of evidence, there would be insufficient evidence to support such a claim.

The challenge for EFSA is to resist the blandishments of the salt reduction lobbyists who claim to "know" that their computer-modeled extrapolations of blood pressure changes due to salt intake levels represent true health outcomes. EFSA should insist not only on good quality science, but on evidence that an intervention (or food) will actually improve HEALTH.