Twelve diet scores compared across menopause. Two separated from the rest.

Weight gain around menopause is one of the most reliably reported experiences of the transition and one of the least well served by evidence. Advice exists in volume; comparative evidence about which dietary pattern actually tracks with less gain during this specific window does not.

A cohort analysis published in JAMA Network Open on May 1 ran the comparison across twelve scoring systems at once [s1].

The design

The study drew on the Nurses' Health Study II, following women over a 12-year period surrounding menopause, with data collected between 1989 and 2019 [s1]. Diet was assessed every four years using validated food frequency questionnaires [s1].

Each participant's diet was scored against a long list of patterns: the plant-based diet index and its healthy and unhealthy variants, the Mediterranean diet, DASH, the Planetary Health Diet Index, low-carbohydrate diets in healthy and unhealthy versions, an empirical dietary inflammatory pattern, the empirical dietary index for hyperinsulinemia, and ultra-processed food intake [s1].

Outcomes were annual change in self-reported body weight and incident obesity [s1].

The analysis included 38,283 women with a mean age of 45.6 years (SD 3.0) [s1]. Across 340,122 person-years of follow-up, 5,214 developed obesity [s1]. Mean weight gain was 0.80 kg per year (SD 1.00) [s1].

That baseline figure deserves its own sentence. In a cohort of health professionals, across the years surrounding menopause, the average was roughly eight kilograms a decade.

Which patterns separated

Models were adjusted for age, race and ethnicity, marital status, income, postmenopausal hormone therapy use, parity, smoking, alcohol, energy intake, physical activity and baseline body mass index [s1].

For weight gain, the largest reduction was associated with the reverse empirical dietary index for hyperinsulinemia — that is, the lowest-insulinemic quintile compared with the highest — at a mean of −0.28 kg per year (95% CI −0.30 to −0.26) [s1]. Against a mean gain of 0.80 kg per year, that is roughly a third of the annual gain.

For incident obesity, the lowest risk was seen for the Planetary Health Diet Index (hazard ratio 0.46; 95% CI 0.42–0.51) and again for the reverse hyperinsulinemia index (hazard ratio 0.51; 95% CI 0.46–0.56) [s1].

What those indices actually contain

Both are empirical scores rather than named diets, which makes them hard to act on unless the foods driving them are known. The paper reports them.

The hyperinsulinemia index showed its largest positive correlations with red or processed meats, sodium, and French fries [s1] — meaning higher intake of those foods pushed the score up, and the protective direction is the reverse.

The Planetary Health Diet Index showed its largest positive correlations with nuts, unsaturated fats, whole grain carbohydrates and vegetable protein [s1].

The authors summarise the pattern as low in red and processed meats, sodium, potatoes and French fries, and rich in nuts, legumes, fruits, vegetables and whole grains [s1].

That is not a novel prescription. What is new is that when twelve scoring systems were tested in the same cohort over the same window, these two separated from the rest — including from DASH and the Mediterranean diet, both of which were in the comparison [s1].

Where to be careful

The hazard ratio of 0.46 for the Planetary Health Diet Index is a large effect for a dietary exposure, and large effects in observational nutrition data usually have company. Women in the top quintile of any healthy-eating score differ from women in the bottom quintile in ways a covariate list does not fully capture, and the direction of that residual confounding is nearly always favourable to the healthy group.

Weight was self-reported [s1]. Diet was measured by food frequency questionnaire [s1], an instrument with known and substantial measurement error. This is a cohort, not a trial: no one was assigned to eat anything, and nothing here demonstrates that changing diet changes weight trajectory through menopause.

The cohort is also drawn from US nurses, a group with more health knowledge and more homogeneous socioeconomic position than the population at large — which narrows both the confounding and the generalisability.

The finding underneath the finding

Weight gain of 0.80 kg per year in a health-literate cohort with above-average access to care [s1] suggests the transition applies a force that dietary pattern modifies rather than cancels. The best performing exposure in this analysis was associated with about a third less annual gain [s1] — real, and not the same as none.

This article is informational and is not medical advice.

Sources

• Optimal Dietary Patterns for Lower Weight Gain and Risk of Obesity Surrounding Menopause — JAMA Network Open, 2026-05-01