Diet, Activity and Genetic Risk in Type 2 Diabetes

A City University of Hong Kong team found that diet and activity patterns were associated with different rates of type 2 diabetes across genetic-risk groups in 56,964 UK Biobank participants. The researchers examined what people ate and how much they moved, then looked at who developed diabetes. They did not assign anyone a new diet or exercise routine, so their headline percentages cannot tell you how much changing your own habits would lower your risk.

People with a lower share of ultra-processed foods in their diets and more moderate-to-vigorous activity were less likely to develop type 2 diabetes, including among participants classified as having higher genetic susceptibility. Inherited risk did not make the observed associations disappear. Nor did it turn the study into a personalized forecast, however tempting a neat percentage might look on a fitness graphic.

The Walk, the Garden and the Wrist Tracker

Wenxin Xu and colleagues examined UK Biobank participants whose average age was 69.5. Dietary information was used to estimate the share of food classified as ultra-processed under the Nova system. Wrist-worn trackers measured moderate-to-vigorous physical activity for one week, and genetic data supplied estimates of susceptibility to type 2 diabetes. During an average 7.8 years of follow-up, 926 participants developed the condition.

Moderate-to-vigorous activity raises heart and breathing rates. The researchers' examples included brisk walking, cycling and hiking, alongside physically demanding housework and gardening. Walking the dog counted as an example, too. The activity measure was not a gym-membership detector. It captured movement, whatever respectable or deeply unglamorous name that movement went by.

The authors reported that participants who met the benchmark of at least 150 minutes of moderate-to-vigorous activity per week were 39% less likely to develop type 2 diabetes than those who did less. They also reported a 7% higher relative risk for every additional 10% of diet, measured by weight, made up of ultra-processed foods. These are comparisons between groups in this cohort. Neither percentage describes the number of cases an individual could expect to avoid by making a change.

The combined diet-and-activity comparison produced the splashiest figures: the authors reported 49% lower relative risk among participants in the high-genetic-susceptibility group, 70% lower in the intermediate group and 59% lower in the low group. For the 70% figure, the comparison was between people at intermediate genetic risk who ate a lower share of ultra-processed foods and met the activity target, and people in that genetic-risk group who ate a higher share and did not meet the target. That figure describes those groups' observed difference, not a 70% discount code for anyone's future diagnosis.

Even the useful activity examples need the same sense of scale as the percentages. A tracker can register a brisk walk or a demanding gardening job during the week someone wears it. It cannot tell researchers whether that person could keep up the same routine throughout the years in which diagnoses were counted. The study asks what patterns were associated with later illness, rather than what would happen if someone started walking the dog more often next Monday.

What One Week and a Food Recall Can Capture

The study followed a large group over time, measured activity with wearables rather than asking people to remember every minute of movement, and linked participants to health records. The researchers also accounted for several other risk factors; the findings were similar when they considered overall diet quality, Marie Spreckley noted in expert comments.

Spreckley identified the dietary method as 24-hour dietary recalls and pointed out that activity was measured for only one week. Participants who ate and moved differently may also have differed in other respects relevant to diabetes risk, even after statistical adjustments. That leaves the effect of deciding to replace foods, increase activity or do both unresolved. It does not erase the observed pattern; it changes the question the study can answer.

Spreckley also said the researchers found no clear evidence that the associations for the individual lifestyle factors differed by genetic risk. Their genetic analyses were limited to people of European ancestry. An association across the groups studied gives researchers reason to examine prevention across genetic-risk categories. It cannot establish the same pattern for every genetic background or show that activity cancels inherited susceptibility.

Public health nutrition researcher Beverley O'Hara offered a further interpretation of the food results: in her view, ultra-processed drinks appeared to drive the observed association with diabetes. She also argued that the 24-hour recall method was not designed to capture ultra-processed-food intake and could misclassify foods by processing level. Her drinks reading is a challenge to treating everything under the ultra-processed label as interchangeable, rather than a finding that every such food has been cleared or condemned. The reported overall percentage cannot serve as a food-by-food shopping list.

Prevention Came Before the Fitness Graphic

Diabetes prevention did not begin with wrist trackers. A 2018 review of lifestyle-prevention research described evidence that structured programs promoting dietary changes and physical activity can prevent or delay type 2 diabetes in higher-risk populations, including people with impaired glucose tolerance. It described approaches delivered through clinical and community settings, from group education and brief counseling to community referrals.

That history gives the new analysis a useful comparison. Prevention interventions examine what happens when people receive a program or support. The UK Biobank analysis compared existing patterns of eating and movement with subsequent diagnoses; its participants were not assigned the combined pattern behind the 70% figure. The intervention evidence helps explain why diet and activity remain serious prevention topics. Its populations and methods differ from this cohort's, so its results cannot convert the cohort's percentages into the effect of following a program.

The review's group education and community referrals also put some daylight between a prevention strategy and an instruction to try harder. A gardening job, a bike ride or a walk with a dog can count as movement in the cohort's activity measure. Whether someone can regularly make room for those activities, or wants a structured program instead, is a practical question a genetic-risk score cannot answer. The older intervention work studies support; the newer cohort adds a view of how activity and diet patterns tracked with diagnoses across the genetic categories it examined.

A 70% figure may fit neatly on a fitness graphic. A brisk walk fits into an actual day. The study measured the association between patterns of days and later diabetes, while prevention programs have tested what happens when people get help changing those patterns.