The association between ultraprocessed food intake and cancer risk Original paper

In this meta-analysis of observational studies, a higher intake of ultraprocessed food was associated with an increased risk of overall cancer.

The association between ultraprocessed food intake and overall cancer risk.

Approximately 430,000 adults with and without cancer.

A systematic review and meta-analysis of 11 observational studies (8 retrospective case-control studies and 3 prospective cohort studies) was performed. Ultraprocessed food was defined according to the NOVA classification system. Examples of ultraprocessed foods include industrially produced breads, cakes and pastries, packaged snacks and candy, commercially made ice cream, breakfast cereals, fruit yogurt, margarine, soft drinks, instant soups, reconstituted meat products, and ready meals.

The studies were conducted in the United States (2 studies) and in Spain, France, Morocco, Iran, Mexico, Chile, Canada, Italy, and South Africa (1 study each). The types of cancer examined in the included studies were colorectal, breast, pancreatic, chronic lymphocytic leukemia, and central nervous system tumors. Most of the studies adjusted for several potential confounders, including BMI and total energy intake.

The meta-analysis indicated that a higher intake of ultraprocessed food was associated with a 12% higher risk of overall cancer in the cohort studies and a 20% higher risk of overall cancer in the case-control studies.

Higher ultraprocessed food intake was associated with the following:

• an increased risk of colorectal cancer (5 of 5 studies), breast cancer (2 of 4 studies, with some evidence indicating that the association was influenced by menopausal status), pancreatic cancer (1 study), and central nervous system tumors (1 study)
• incident cases (diagnosis within less than 1 year of data collection) of chronic lymphocytic leukemia, but not prevalent cases (diagnosis 1–3 years prior to data collection), in 1 study.

Higher ultraprocessed food intake was not associated with prostate cancer risk (3 of 3 studies).

A limitation of this study is that the food frequency questionnaires used to assess dietary intake in most studies were not specifically designed for use with the NOVA classification system, and consequently, several studies reported difficulties in classifying food items due to insufficient information available in the questionnaire. As such, misclassification of foods may have influenced the results.

Additionally, as evidenced by the examples provided above, the ultraprocessed food category is highly heterogeneous, with the degree of potential harmfulness likely to vary considerably between the included foods. These issues limit our confidence in the findings.