The Slow Retreat of Salt

ASalt is not a modern indulgence but a physiological necessity: sodium governs the volume of fluid the body retains and the electrical signalling of nerve and muscle. The quantity required, however, is small. International advice, unchanged since the 1990s, puts the safe ceiling at five grams of salt a day, roughly a level teaspoon, while adults in Vantry were consuming 9.6 grams in 2003. What made that figure so resistant to persuasion was its origin. Dietary surveys commissioned by the Vantry health ministry found that 76 per cent of the salt in the national diet had been added before food reached a domestic kitchen, and that only 15 per cent came from the salt cellar or the cooking pot. Bread alone supplied 18 per cent. A decade of leaflets urging shoppers to season more sparingly had moved consumption by an amount too small to measure. Whatever was to be done would have to be done by manufacturers.

BThe Vantry programme, launched in 2004, therefore addressed itself to producers rather than to the public. A National Salt Board set voluntary targets for 85 categories of processed food, revised every three years, on a principle borrowed from sensory science: a single reduction of five per cent went undetected by taste panels, whereas the same total cut imposed in one step was rejected outright. By 2015 the salt in an average loaf had fallen by 22 per cent and that in breakfast cereals by 31 per cent; measured intake stood at 8.1 grams. Sixty-two of the 85 categories met their targets, while sauces and cured meats missed theirs at every revision. Retailers' own-brand ranges, which by then accounted for two loaves in every five sold in Vantry, moved fastest of all. Deaths from stroke fell by 34 per cent across the same eleven years, a figure quoted enthusiastically abroad. It is also uninterpretable on its own: statin prescribing trebled and smoking declined sharply over the period, and the programme was never accompanied by a comparison population.

CMiriam Osei-Bonsu, who leads the diet policy group at the Ravensworth Institute, regards the work as unfinished. Pooling 34 controlled trials, she reports that removing 4.4 grams of salt a day lowers systolic blood pressure by 4.1 millimetres of mercury in people already diagnosed with hypertension, and by 1.6 in those without it. Feeding those coefficients into a simulation of the Vantry population, she calculates that a further one-gram cut would avert some 4,300 deaths a year at a cost of roughly fifteen pence a head. She is candid about the status of the number: these are modelled outcomes, she writes, and no trial has ever randomised a whole country. Her defence is that the alternative to modelling is not certainty but inaction, and that reformulation carries no plausible harm even if the projected benefit turns out to have been overstated.

DAnders Lindqvist, an epidemiologist at the Calloway School of Public Health, doubts the lower half of the curve. His observational cohort of 28,000 adults, followed for eleven years, found cardiovascular risk lowest among participants excreting the equivalent of seven to twelve grams of salt a day, and higher at both extremes: a J-shaped relationship rather than a straight line. He does not argue that very high intakes are safe. His claim is narrower, that below roughly seven grams the evidence of further benefit thins out. Osei-Bonsu's rebuttal has two parts. The upturn at the low end, she argues, is reverse causation, since people who are already ill eat less, so that their low readings follow from disease rather than produce it. And Lindqvist's estimate rests on a single urine collection taken at entry to the study, a method known to misclassify habitual intake.

EWhether targets should be voluntary at all divides administrations more than it divides scientists. Kesland, Vantry's neighbour, made limits compulsory in 2018, but wrote them for two categories only, bread and cured meat. Compliance reached 94 per cent within two years, and yet national intake fell by 0.4 grams against the 1.1 grams the ministry's own model had predicted. The audit that followed found that salt in the uncapped categories, among them ready meals, soups and cheese, had risen by 9 per cent over the same period, enough to swallow most of the expected gain. Whether producers rebalanced their recipes deliberately or merely followed a drift towards stronger flavours elsewhere, the auditors declined to say. Vantry's board drew the obvious lesson and now revises its entire list of categories simultaneously.

FThe most striking result of the decade came from neither camp. In Ardmore province, 21,000 villagers registered at 148 rural clinics were supplied with a substitute in which a quarter of the sodium chloride had been replaced by potassium chloride; over four years the stroke rate in those villages ran 12 per cent below the rate in villages using ordinary salt. The finding does not travel without care, because potassium accumulates dangerously in people whose kidneys are failing, and the trial excluded such patients by screening them beforehand. Ardmore has since extended the scheme to nine further provinces, this time with no comparison villages at all. On one point the two sides now agree. Blood pressure is a proxy, and any programme should be judged on trials long enough to count deaths; until such trials exist, Lindqvist accepts that reformulation is cheap and probably harmless, while Osei-Bonsu accepts that the intake data underpinning every model in the field remain poor.