The Gale readResearch

Diet and hip fracture after menopause: 60 of 99 factors linked

A 19.7-year follow-up of 541,887 UK postmenopausal women found chicken, vegetables and pasta tied to fewer hip fractures and chips and fizzy drinks to more, and the authors say causation is not settled.

By Gale Staff · October 14, 2026 · European Journal of Epidemiology

Not reviewed by a clinician.

Illustration: Menopause & midlife

The short answer

In a UK cohort of 541,887 postmenopausal women followed for an average of 19.7 years, 60 of 99 dietary factors were associated with hip fracture after correction for multiple testing, with foods usually called healthy tending toward lower risk and foods usually called unhealthy toward higher risk. Chicken (hazard ratio 0.84 per 20 g/day), vegetables (0.88 per 100 g/day) and pasta (0.87 per 20 g/day) sat on the lower side, chips (1.14 per 25 g/day) and fizzy drinks (1.18 per 50 g/day) on the higher side, but this is an observational study, and the authors state that it remains to be shown whether the associations are causal or reflect overall diet quality and other lifestyle differences.

Who to see

Gynecologist or menopause specialist

After menopause, foods usually called healthy tended toward lower hip fracture risk and fizzy drinks toward higher; a menopause specialist can talk through your bone health.

Find care →

What a long diet-and-bones question looks like from the dinner table

A woman in her late fifties is clearing a plate of chicken and vegetables and wondering, not for the first time, whether any of this matters for her bones twenty years from now. Hip fracture is a serious outcome of later life, and the advice around food and bone health is usually a short list of nutrients that most people cannot connect to an actual plate. The question that follows any headline about diet and fracture is a practical one: which everyday foods show up in the data, how large are the differences, and how far can anyone trust them.

A paper in the European Journal of Epidemiology, published online on 12 October 2026 by Yang Tao, Timothy Key, Gillian Reeves, Sarah Floud, Keren Papier and Tammy Tong, takes that question on at unusual scale. Earlier research on diet and hip fracture, the authors note, had focused on selected foods and nutrients. This study instead looked at 99 dietary factors together, an approach called a diet-wide association study, meaning every food and nutrient is tested against the same outcome with the same method.

What the researchers did

The data come from the Million Women Study in the United Kingdom. The analysis included 541,887 postmenopausal women. Their usual diet was assessed with a food frequency questionnaire, a checklist of how often foods are eaten, completed in 2000 to 2004, and the authors report that it is a validated instrument and that the intakes were calibrated using a 24-hour dietary recall, which is a more detailed account of what a person ate the previous day.

Hip fractures were identified through linked hospital admission records, so a fracture counted when it led to a hospital stay. The team used Cox regression, a standard method for comparing how quickly an event occurs in groups with different exposures. The result is a hazard ratio: a value of 0.84 means a 16 percent lower rate of the event at a given moment for each stated increase in intake, and a value of 1.14 means a 14 percent higher rate. Each result comes with a 95 percent confidence interval, the range of values reasonably compatible with the data.

Testing 99 factors invites false positives, so the authors applied a correction for multiple testing before deciding which associations to report. They also checked whether results held in subgroups and when the first five years of follow-up were excluded, a common way to reduce the chance that undiagnosed illness early on was already shaping what women ate.

What they found

After an average of 19.7 years of follow-up, 27,318 women had a hip fracture, which is about 5 percent of the 541,887 women studied. That share is simple arithmetic on the two published totals and not a figure from the paper, and it ignores the different lengths of follow-up, so it is only a sense of scale. In models that adjusted for other factors, 60 of the 99 dietary factors were associated with hip fracture risk after multiple testing correction.

The five foods with the strongest statistical evidence, ranked by p-value, were chicken, vegetables, pasta, chips and fizzy drinks. Chicken had a hazard ratio of 0.84 (95 percent confidence interval 0.80 to 0.87) per 20 g per day. Vegetables came in at 0.88 (0.85 to 0.91) per 100 g per day, and pasta at 0.87 (0.83 to 0.91) per 20 g per day. In the other direction, chips had a hazard ratio of 1.14 (1.10 to 1.18) per 25 g per day, and fizzy drinks 1.18 (1.12 to 1.24) per 50 g per day.

The top five nutrients followed a similar pattern. Protein had a hazard ratio of 0.79 (0.75 to 0.84) per 15 g per day, zinc 0.81 (0.77 to 0.86) per 2 mg per day, carotene 0.91 (0.88 to 0.94) per 1000 micrograms per day, fibre 0.89 (0.85 to 0.92) per 5 g per day, and niacin 0.82 (0.77 to 0.87) per 10 mg per day. All five sit below 1, meaning higher intake went with a lower rate of hip fracture. The authors report that results were consistent across subgroups and when the first five years were dropped, and they sum up the pattern as many healthy foods associated with lower risk and unhealthy foods with higher risk.

The caveat that carries the paper

The authors close with a plain statement: further research is needed to assess whether these associations are causal or reflect overall diet quality or other lifestyle confounding. That sentence deserves to be read as the main limit and not as a formality. In observational data, women who eat more vegetables can differ from women who eat more chips in other ways that also bear on bones, and statistical adjustment can reduce but not remove those differences. The authors name lifestyle confounding as one possibility they could not rule out.

There is a second reason for caution. Foods and nutrients overlap. Chicken is a source of protein and zinc, and vegetables a source of carotene and fibre, so the 60 associations are probably not 60 independent findings. A list of 60 factors partly reflects one underlying contrast, between eating patterns that are generally rated as healthier and those that are not. That is an interpretation of the pattern, not something the abstract tests.

Finally, the abstract describes diet as measured once, in 2000 to 2004, and the follow-up runs about twenty years. It does not say whether diets were reassessed, and eating habits change over two decades. Measurement error in a food questionnaire is also unavoidable; the calibration with a 24-hour recall addresses part of that and does not remove it. The abstract does not report absolute risks, meaning the actual number of fractures per thousand women in the highest and lowest intake groups, so the size of any practical difference cannot be taken from the hazard ratios alone.

Why it matters even without a verdict

Hip fracture is a serious event that, by the authors' own method, is counted only when it reaches a hospital, and patterns that might be modifiable are worth mapping. The contribution of this paper is breadth and consistency: one cohort, one method and one outcome applied to 99 factors, with a large enough count of fractures, 27,318, to give narrow confidence intervals. That makes it a useful map of which dietary signals are strongest in this population and a sensible source of hypotheses for trials or for studies that can separate single foods from whole diets.

What it cannot do is tell any individual woman that adding a particular food will lower her own risk. The associations describe groups over time, in one country, among women who completed a questionnaire in the early 2000s. How much of the difference food itself explains is the open question the authors leave in their final sentence.

What this study can't tell you

  • Whether any of these foods or nutrients cause a change in hip fracture risk: this is an observational cohort, and the authors say it remains to be shown whether the associations are causal or reflect overall diet quality or other lifestyle confounding.
  • How large the absolute differences are: the abstract reports hazard ratios and confidence intervals, not fracture rates per thousand women in the highest and lowest intake groups, so a practical size of effect cannot be read from it.
  • Whether the 60 associations are 60 separate findings: foods and nutrients overlap (protein, zinc, fibre and carotene come from foods that also appear on the list), so the results likely reflect fewer underlying contrasts, and the abstract does not test that.
  • Whether diet changed over the follow-up: the abstract describes dietary assessment in 2000 to 2004 and an average follow-up of 19.7 years, and does not say whether intake was measured again.
  • Whether the results apply beyond postmenopausal women in the UK Million Women Study, such as men, younger women, or populations with different diets and fracture rates.
  • Which of the roughly 50 other associated factors beyond the ten named in the abstract matter, since only the top five foods and top five nutrients are reported there with numbers.

The Gale read

This is a careful, very large piece of descriptive work whose most useful property is that its authors state its limit in their own last sentence. The strengths are real: 541,887 women, 27,318 hospital-recorded hip fractures, one method applied to 99 factors with a correction for multiple testing, and results that held in subgroups and after dropping the first five years of follow-up. The claim worth carrying is the narrow one, that in this cohort a large number of dietary factors line up with hip fracture risk in a direction that tracks conventional ideas of a healthy diet. The broader claim, that eating more of the lower-risk foods would prevent fractures, is not what the study measured, and the authors do not make it. Because the foods and nutrients overlap, and because lifestyle differences travel with diet, the paper is best read as a ranked map of signals for the next study to test, not as a list of things to eat or avoid. The fracture rates behind the hazard ratios would turn this from a map into something closer to a measure of consequence, and the abstract does not carry them.

Common questions

what foods are linked to lower hip fracture risk in older women

In a UK study of 541,887 postmenopausal women, the foods with the strongest evidence for lower hip fracture risk were chicken (hazard ratio 0.84 per 20 g/day), vegetables (0.88 per 100 g/day) and pasta (0.87 per 20 g/day). Higher intakes of protein, zinc, fibre, carotene and niacin also went with lower rates. These are associations from an observational study, and the authors say they may reflect overall diet quality or other lifestyle differences.

do chips and fizzy drinks raise the risk of hip fracture

They were associated with a higher rate in this cohort: a hazard ratio of 1.14 per 25 g/day for chips and 1.18 per 50 g/day for fizzy drinks. That means a 14 percent and an 18 percent higher rate of hip fracture for each stated increase in intake. Because the study is observational, it cannot show that the foods themselves caused fractures, and the authors list lifestyle confounding as an unresolved possibility.

does protein help prevent hip fractures

Higher protein intake was associated with a lower rate of hip fracture, with a hazard ratio of 0.79 (95 percent confidence interval 0.75 to 0.84) per 15 g/day. That is the lowest ratio of the five nutrients named in the abstract, though each is measured per a different unit. The design shows a link in women who reported their diet in 2000 to 2004, and it cannot show that adding protein would change anyone's fracture risk.

how many postmenopausal women get hip fractures

In this study 27,318 of 541,887 postmenopausal women had a hip fracture over an average of 19.7 years of follow-up, which is roughly 5 percent as a crude share. That figure is simple arithmetic on the two totals and ignores differences in follow-up time, so it gives a sense of scale and not a risk estimate for any particular woman.

Sources

  1. 1.Tao Y, Key TJ, Reeves GK, Floud S, Papier K, Tong TYN. Diet-wide association study (DWAS) of hip fracture risk in UK postmenopausal women: a prospective cohort study of 541,887 women with 27,318 incident cases over ~20 years. European Journal of Epidemiology. 2026 (published online 12 October 2026). doi:10.1007/s10654-026-01466-3 link

Read the study →

1 sources, numbered by first appearance. General health information, not medical advice. AI-assisted editorial content — every citation independently verified. Editorial policy

More from Gale Staff

The Latest Health Research →

Published October 14, 2026 · Gale Staff

All Gale Staff analyses →