Reading the studies together is useful because their numbers describe different outcomes. The first follows disease events in cohorts. The second calculates a risk score after controlled dietary periods. The distinction determines what a headline percentage can actually tell us.
The plant-based study measured more than heart disease
The Lancet Healthy Longevity analysis by Córdova and colleagues included 407,618 participants: 226,324 in EPIC and 181,294 in UK Biobank. Its central outcome was cancer–cardiometabolic multimorbidity, defined around developing at least two diseases among cancer, cardiovascular disease and type 2 diabetes. That is more specific than “chronic disease,” and different from heart disease alone.
Each ten-point increase in the healthful plant-based diet index, or hPDI, was associated with a hazard ratio of 0.89 in EPIC and 0.81 in UK Biobank for multimorbidity. The reported 95% confidence intervals were 0.83–0.96 and 0.76–0.86. These are the source of the roughly 11% and 19% lower relative hazards. They are not percentage-point reductions in an individual’s absolute risk.
Median follow-up was 10.9 and 11.4 years, respectively. The study was observational: participants were not randomly assigned to maintain different hPDI scores. Adjustment for measured differences strengthens the analysis but cannot eliminate every alternative explanation. It supports an association, not a promise that increasing a score by ten points produces the same benefit for every reader.
Plant-based describes quality as well as ingredients
The important practical distinction is between a diet’s label and the foods that earn it. Removing an animal ingredient does not, on its own, establish the nutritional quality of its replacement. “Plant-based” is too broad to serve as an explanation for every food pattern or every health outcome.
Nor does the way a food is manufactured determine the health effects of an eating pattern. Our review of cultivated meat addresses production assumptions and product-specific oversight, rather than claiming that a manufacturing category establishes a cardiovascular benefit.
For an established description of a heart-healthy pattern, the US National Heart, Lung, and Blood Institute’s DASH guidance includes vegetables, fruit, whole grains, beans and nuts, alongside options such as low-fat dairy, fish and poultry. It also recommends limiting saturated fat and sugar-sweetened drinks. DASH does not require a completely vegan diet or special branded foods.
The DASH-Sodium analysis used a calculated risk score
The 2025 American Journal of Preventive Cardiology paper analysed 412 participants in the DASH-Sodium trial. They were assigned to a DASH or control eating pattern and consumed three sodium levels in random order within their dietary arm. Each feeding period lasted 30 days. The analysis used the Pooled Cohort Equations to estimate ten-year atherosclerotic cardiovascular disease, or ASCVD, risk.
Compared with the high-sodium control diet, the combined low-sodium DASH condition produced about a 14.1% relative reduction in the estimated score. The paper reports an absolute difference of approximately 0.35 percentage points. That is not a reduction of 14.1 percentage points, and it is not an observed 14.1% fall in heart attacks over ten years.
Sodium reduction and the DASH pattern each contributed to the result. However, the score is a prediction derived from risk factors; a favourable change is not equivalent to measuring clinical events throughout the prediction period. The study concerns adults with elevated blood pressure and no prior cardiovascular disease, and does not establish identical effects in every patient group.
Why the percentages cannot be added or ranked
The 11% and 19% figures concern hazards of multimorbidity in two long-term cohorts. The 14.1% figure concerns a relative change in a calculated ASCVD risk score under one dietary comparison. Different populations, outcomes and methods are involved. Adding the figures would create a number that neither study measured.
The absolute and relative descriptions of the DASH result also answer different questions. A difference expressed in percentage points describes the change in the score’s level. A relative percentage expresses that change in proportion to a comparison value. Reporting both prevents a modest absolute change from sounding like a much larger fall in the number of people affected.
What a reader can reasonably use
These results support paying attention to an overall eating pattern rather than treating a list of “heart foods” as a prescription. For practical food choices and the role of sodium, use the DASH eating-plan guidance and its serving information. The number of servings depends on energy needs; the research percentages are not a method for calculating a personal menu or deciding whether treatment can be stopped.
For tools intended to support a change in eating behaviour, our review of digital weight-loss tools examines what monitoring adds to a programme that also includes professional support. A glucose trace, weight change and long-term cardiovascular outcome remain different measures.
If a change needs to fit an existing medical condition or treatment plan, that requires individual advice. The useful reading habit is simpler: ask which people were studied, what was measured, over what period, and whether the figure is a relative effect, an absolute difference or a predicted score. Here those questions turn an overbroad prevention headline into two distinct, informative findings.
