Results from a large-scale analysis show that following a vegetarian or pescatarian (fish-eating) diet could significantly reduce the risk of developing cancer – but even limiting red and processed meat to five meals a week or less may also have a benefit.
Cancer is a leading cause of death worldwide, responsible for approximately 10 million deaths in 2020, with breast, colorectal and prostate among the most commonly diagnosed types. However, it has been estimated that around 40% of cancer cases may be caused by modifiable factors, and thus could in theory be prevented. Evidence from previous large cohort studies has indicated that vegetarians may have a lower risk of developing all types of cancer compared to regular meat-eaters, however it was unclear whether this varied by different cancer sites. It was also unknown whether cancer risk was different for people who have a reduced intake of meat, or who did not eat meat but ate fish (fish-eaters or pescatarians).
Researchers from Oxford Population Health’s Cancer Epidemiology Unit (CEU) analysed data from over 472,000 participants in the UK Biobank to investigate the association between diet and cancer risk. The results have been published today in BMC Medicine.
British participants (aged between 40-69 years) were recruited to the UK Biobank Study between 2006 and 2010, and completed questionnaires about how often they ate meat and fish. The participants were then categorised into four dietary groups: regular meat-eaters (eating meat more than five times a week); low meat-eaters (eating meat five times a week or less); fish-eaters (do not eat meat, but eat fish); and vegetarians (do not eat meat or fish).
The final study sample, who had no history of cancer at recruitment, included 247,571 (52.4%) regular meat-eaters; 205,385 (43.5%) low meat-eaters; 10,696 (2.3%) fish-eaters; and 8,685 (1.8%) vegetarians (which also included 446 vegans). The participants were followed up for an average of eleven years through linkage to public health records.
- Over the follow-up period, there were 54,961 new cancer diagnoses, including 5,882 colorectal, 7,537 postmenopausal breast, and 9,501 prostate cancers.
- Compared with regular meat-eaters, the risk of developing any type of cancer was lower in low meat-eaters (2% less), fish-eaters (10% less), and vegetarians (14% less). This means that the absolute reduction in cancer diagnoses for vegetarians was 13 fewer per 1,000 people over ten years, in comparison to regular meat-eaters.
- The risk of postmenopausal breast cancer was significantly reduced in vegetarians (18% less), compared with regular meat-eaters. However, additional analyses indicated that most of this reduction in risk was due to vegetarians having a lower average body mass, compared with regular meat-eaters.
- The risk of prostate cancer was significantly reduced in both vegetarians (31% less) and fish-eaters (20%), compared with regular meat-eaters. This equates to 11 and 7 fewer diagnoses per 1,000 people over ten years respectively, in comparison to regular meat-eaters.
- In men, compared with regular meat-eaters, the risk of colorectal cancer was lower in low meat-eaters (11% less), fish-eaters (31% less), and vegetarians (43%). However, there was no apparent difference in risk for women for any of these dietary groups.
The researchers also found no significant evidence that these results were driven by differences in testosterone and insulin-like growth factor-1 (IGF-1), two hormones that have been linked to cancer, although IGF-1 levels were slightly lower in vegetarians.
Lead author for the study Cody Watling, a DPhil student within CEU, said: ‘Our study found that being a low meat-eater, fish-eater, or vegetarian was associated with a significantly lower risk of all cancer. But it is not yet clear whether these differences are due to dietary factors or non-dietary effects, as differences in use of health services, including cancer screening, may also have influenced our findings. Future research assessing cancer risk in cohorts with larger numbers of vegetarians is needed to explore in greater detail the possible explanations for these observed differences.’