Frequent High-Sugar Foods Linked to 28% Higher Breast Cancer Risk in Korean Cohort, Fruit Sugar Not

Between 2004 and 2013, researchers in South Korea enrolled 169,718 adults aged 40 to 69 at 38 centers for a community-based project called the Health Examinees study, or HEXA. A recent analysis from Seoul National University, published in the European Journal of Nutrition, draws on a subset of that cohort to ask a narrow question: does breast cancer risk track with how much sugar women eat, or with where the sugar comes from?

The subset, called HEXA-G, held 92,314 women before exclusions. Women were dropped if they had not consented to registry linkage, had incomplete dietary questionnaires, reported implausible energy intake, or had any cancer before enrolling. Anyone diagnosed with breast cancer within two years of enrollment was also removed, to reduce the chance that undiagnosed disease at baseline would skew results. That left 71,264 women.

Diet was measured with a semi-quantitative questionnaire covering 106 foods. Participants reported how often they had eaten each item over the previous year and in what portion size. The team built a sugar database for all 106 items from national and international food composition tables and nutrition labels, then defined seven high-sugar groups: carbonated drinks, Korean traditional sugar-based tea, cakes and cookies, sweets, sugary cereals, ice cream, and a category of jam and butter that also includes honey and margarine. Breast cancer cases through December 31, 2018, were identified through the Korea Central Cancer Registry. The models adjusted for smoking, alcohol, education, exercise, reproductive history, family history, BMI, and energy intake.

Over an average of 9.2 years of follow-up, 713 women were newly diagnosed. Mean sugar intake was 44 grams a day, and fruit supplied about 43% of it. Women who ate high-sugar foods more than twice a week had a 28% higher risk than those who ate them once a month or less. Measured by sugar content rather than servings, the top quarter of consumers had a 28% higher risk than the bottom quarter.

Two foods stood out. Women who ate ice cream at least weekly had a 26% higher risk than those who did not, though that estimate was borderline in statistical terms. Eating sugary cereal at least once every two months was linked to a 34% higher risk than never eating it, but the authors noted that cereal consumption in the cohort was low. No significant associations appeared for carbonated drinks, cakes and cookies, sweets, jam and butter, or traditional Korean tea. Total sugar, fruit sugar, fruit servings, carbohydrates, and fat also showed no significant link.

Subgroups added some texture. The highest-versus-lowest comparison was significant among women under 50, but the trend across categories was not, and the age interaction was not significant, so the study did not establish that age groups differ. A significant association also appeared among women who drank alcohol. Interaction tests hinted at differences by alcohol use, BMI, and physical activity that the authors say need further investigation.

The authors read the pattern as a sign that the source and context of sugar may matter alongside quantity. Fruit contributed the most sugar yet showed no association, which points attention toward processed, sugar-rich foods. But the study could not separate added from naturally occurring sugars, so it cannot say whether sugar itself or other components of these foods explain the results.

The limitations are substantial. Diet was assessed once, at baseline. There was no data on cancer subtype or stage, or on menopausal status at diagnosis. Family history data were missing in places, excluding early diagnoses may have introduced bias, and residual confounding is possible. The design is observational, so it cannot show cause, and the proposed biological mechanisms were not tested. The authors call for future work using repeated dietary measures, subtype data, and menopausal status at diagnosis.