a close up of a plate of food with veggies

Can Diet Stop Menopause Weight Gain?

A 12-year study of 38,283 US women tied plant-forward diets to slightly less weight gain at menopause — an association, not proof diet stops it.

In 12 years of follow-up across 38,283 US women, diets highest in plants or lower in sugar tracked with slightly less yearly weight gain and lower obesity risk.

In a 12-year follow-up of 38,283 US women around menopause, those whose usual diets scored highest on a planetary-health or low-insulinemic pattern gained slightly less weight each year and were less likely to become obese. The largest yearly weight difference was 0.28 kg (about 0.6 lb) — an association, not proof the diets stopped the weight gain of menopause.

What this means

In 38,283 US women followed for 12 years around menopause, diets that scored highest on a planetary-health (mostly plant-based) or low-insulinemic pattern were associated with slightly less weight gain — at most about 0.28 kg (0.6 lb) a year — and with a lower likelihood of becoming obese, while patterns higher in red and processed meat, sodium, potatoes and French fries tracked with more weight gain. The study is observational: the women were not assigned to any diet, and those who eat healthier patterns may differ in other ways, so it cannot show the diets caused the difference. The absolute difference is small, and because the cohort was predominantly White female health professionals who reported their own diet and weight, the findings may not generalize to other groups.

The strongest signals were modest

The study followed 38,283 women in the Nurses’ Health Study II for 12 years surrounding menopause, from 1989 to 2019 — 340,122 person-years in all. The women were, on average, 45.6 years old at the start. Their diets were measured every four years with a validated food frequency questionnaire, and the researchers computed 11 diet-pattern scores plus ultra-processed food intake, then tested each against annual weight change and new cases of obesity.

The women gained weight on average: 0.80 kg (about 1.8 lb) a year. Against that backdrop, the patterns that scored higher on plant-forward, low-insulinemic ways of eating were associated with slightly less gain. The single largest difference was for the reverse empirical dietary index for hyperinsulinemia (a lower-insulinemic pattern): women in the top fifth gained 0.28 kg (about 0.6 lb) less per year than women in the bottom fifth, with a 95% confidence interval of −0.30 to −0.26 kg per year. Across all the healthy patterns that reached significance, the yearly differences ran from that 0.28 kg down to 0.06 kg (about 0.1 lb) for the healthy low-carbohydrate pattern.

For obesity, the planetary health diet index carried the lowest risk: the top fifth had a hazard ratio of 0.46 (95% CI 0.42 to 0.51) compared with the bottom fifth. The lower-insulinemic pattern came in at 0.51 (95% CI 0.46 to 0.56), and DASH was the weakest significant pattern at 0.85 (95% CI 0.77 to 0.93). Unhealthy scores — the low-carbohydrate, unhealthy low-carbohydrate, inflammatory and hyperinsulinemic patterns, and ultra-processed food intake — pointed the other way, tracking with more weight gain and higher obesity risk. The researchers controlled for multiple comparisons using the Bonferroni method within each outcome family.

A large hazard ratio, a small absolute difference

The two numbers need to be read together. A hazard ratio of 0.46 is large for a single dietary pattern in an observational study, and associations this strong in observational data are the kind that warrant caution about residual confounding — the possibility that the women who ate this way differed in other ways the analysis did not fully capture. The absolute difference is the more honest headline: under 0.3 kg a year at the very best, against an average yearly gain of 0.80 kg.

Food groups pointed the same direction

The same split appeared at the food level. Red and processed meats, poultry, animal monounsaturated fat and animal protein, low-energy drinks, sodium, and French fries and potatoes were associated with more weight gain over the follow-up, while nuts, legumes, unsaturated fats, fruit carbohydrate and whole-grain carbohydrate were associated with less. These are the building blocks of the pattern scores, which is why several indices overlap.

The result held up in sensitivity checks: averaging diet within each four-year interval, or adjusting for physical activity, did not change it, and it was consistent across pre- and post-menopausal women. One association — between the inflammatory pattern and weight change — was significantly larger among women using postmenopausal hormone therapy.

What the study cannot show

This is a prospective observational cohort, not a randomized trial. The women were not assigned to a diet, so the study can show that eating patterns and weight moved together but not that the patterns caused the difference. Diet and weight were both self-reported (weight validated against technician measures). The cohort was predominantly White US female health professionals, a group the authors note limits how far the findings travel. The study also has no body-composition data, so it cannot separate fat from lean mass; no menopausal-symptom or temporally aligned sex-hormone measures; and too few clinical events during menopause — such as type 2 diabetes, heart disease or cancer — for a robust analysis of those outcomes.

Where this sits

Most evidence on diet and menopausal weight is observational, and large comparisons of many dietary patterns at once are scarce, and randomized trials of dietary patterns across menopause are largely absent. This study extends earlier work in the study family’s parent cohorts — the original Nurses’ Health Study and the Health Professionals Follow-up Study — finding that diets with lower insulin-stimulating and inflammatory potential tracked with less weight gain (Tabung et al., Journal of Nutrition 2019), and it lines up with companion work from the same group showing that the quality of a low-carbohydrate diet, not just its macronutrient share, tracked with weight change (Liu et al., JAMA Network Open 2023). What remains unsettled is the part the press framing tends to skip: no randomized trial has tested whether adopting a planetary-health or low-insulinemic pattern across menopause actually leads women to gain less weight. For a related look at what supports weight maintenance rather than menopausal weight change, see steps-weight-maintenance-meta-analysis.

Funding, interests and access

The Nurses’ Health Study II is funded by National Institutes of Health grants U01 CA176726 and U01 HL145386; Dr Haslam by NIH K01 DK136968; and the work also drew on T32 CA 009001 and the Julie Wilcox Women’s Health and Nutrition Novel Research Fund. The funders had no role in the design, analysis or decision to publish. Disclosed interests outside the work include NIH grants for Dr Manson, grants from the Almond Board of California for Drs Sun and Bhupathiraju, NIH grants for Dr Willett during the study, and a grant from Analysis Group for Dr Hu. The full text was available to us in open access at press time.

Terms explained

  • Observational cohort vs randomized trial: a cohort follows people going about their usual lives and records what happens, so it can find that two things track together but not prove one caused the other. A randomized trial assigns people to a diet or a comparison, which is what makes cause-and-effect inferences possible.
  • Food frequency questionnaire (FFQ): a survey asking how often a person ate each of a list of foods over a set period. It captures usual diet but relies on memory and estimation, so it carries error.
  • Hazard ratio and 95% confidence interval: a hazard ratio compares the rate of a new outcome in one group with another over time; 0.46 means the higher-scoring group had about 46% of the comparison group’s rate. The 95% confidence interval is the range of values consistent with the data — here 0.42 to 0.51 — and a range that does not cross 1.0 indicates the difference is unlikely to be chance.
  • Planetary health diet index (PHDI): a score that rates a diet higher when it emphasizes plant foods such as vegetables, whole grains, nuts and legumes and lower when it leans on red and processed meat and other animal foods.
  • Empirical dietary index for hyperinsulinemia (EDIH): a score, built from foods most linked to blood-insulin response, that rates a diet higher when it is more likely to stimulate insulin. A lower EDIH score — the “reverse” pattern in this study — describes a less insulin-stimulating diet.
  • Perimenopause and menopause: the transition years when ovarian hormone production declines, and the point 12 months after a final period. Weight and body composition often shift during this window.

Sources

  1. Primary source: Xia T, Haslam DE, Eliassen AH, Manson JE, Sun Q, Willett WC, Bhupathiraju SN, Zhang C, Hu FB. “Optimal Dietary Patterns for Lower Weight Gain and Risk of Obesity Surrounding Menopause.” JAMA Network Open 2026;9(5):e2613102; published online May 2026. DOI: 10.1001/jamanetworkopen.2026.13102 · PMID: 42160056 · PMCID: PMC13191392. Peer-reviewed prospective cohort study (observational; association only). Full text read (open access, CC BY).
  2. Landscape — the predecessor finding (abstract-level access): Tabung FK, et al. “Long-Term Change in both Dietary Insulinemic and Inflammatory Potential Is Associated with Weight Gain in Adult Women and Men.” Journal of Nutrition 2019;149(5):804–815. DOI: 10.1093/jn/nxy319 · PMID: 31004153 · PMCID: PMC6499102. The prior NHS/HPFS finding (women in the original Nurses’ Health Study and men in the Health Professionals Follow-up Study) that this menopause-specific study extends.
  3. Landscape — companion work (abstract-level access): Liu B, et al. “Low-Carbohydrate Diet Macronutrient Quality and Weight Change.” JAMA Network Open 2023. DOI: 10.1001/jamanetworkopen.2023.49552 · PMID: 38150249 · PMCID: PMC10753393. Same research group; diet quality — not just macronutrient share — tracks weight change. An erratum correcting Figure 2 was published in 2024 (JAMA Netw Open 2024;7(2):e240829); the conclusion cited here is unaffected, as it comes from the study’s reported results rather than Figure 2.
  4. Landscape — the field review (abstract-level access): “A comprehensive evaluation of predictors of obesity in women during the perimenopausal period: a systematic review and narrative synthesis.” Diabetes & Metabolic Syndrome 2024. DOI: 10.1016/j.dsx.2023.102933 · PMID: 38181722. Notes that randomized trial evidence on dietary patterns across menopause is scarce.
  5. Press signal (discovery and framing only, not evidence): “Healthy dietary patterns can help control menopause-related weight gain.” Harvard T.H. Chan School of Public Health, 2026. https://hsph.harvard.edu/news/healthy-dietary-patterns-can-help-control-menopause-related-weight-gain/

Related Articles