(View plain language abstract)
Background and aim
The AHS-2 is a cohort comprising a special North American population (N=96,000) of whom half are vegetarian, and 25% are Black (African-American and West Indian). The average age at enrolment was 62 years and 65% of participants were female.
The main goals of this study were to capitalise on the unusual dietary characteristics of this population, of whom about half were vegetarian (including 8% vegan), and half of whom ate soy at Asian levels, and to relate the risk of breast, colorectal, and prostate cancers to different dietary habits taking advantage of the wide range of dietary exposure in this population.
This feature adds power and reduces the biasing effects of measurement error.
Methods
Participants were enrolled in the study by promotional efforts among 4,000 congregations across North America. Enrolment was concluded when they completed and mailed back a long questionnaire about dietary habits and other aspects of lifestyle and medical history.
New cancers were found by matching AHS-2 data with each state cancer registry. A major challenge in the US is to detect all incident cancer cases from a nationwide cohort. The WCRF grant has been most helpful in providing supplementary funds to enable us to successfully achieve the goal of matching with all US cancer registries (except in Maine), a task not previously accomplished by a research group.
Based on this, and again aided by some support from the WCRF grant, we have been able to pursue analyses and the production of manuscripts to address the main study goals.
Results
Results at present are as follows:
- As compared to non-vegetarians there is little difference in the cancer mortality of vegetarians; however the incidence of gastrointestinal cancers in vegetarians is reduced, as is the risk of cancers of female genital organs in females;
- Vegetarians have a significant 20% reduction in colorectal cancer, and it appeared that this was nearly 50% in pesco-vegetarians, although this difference from other vegetarians was not statistically significant;
- Vegans, as compared to non-vegetarians and other vegetarians, appear to have a significant 33% reduction in the risk of total prostate cancer, a reduction that is maintained in aggressive/advanced incident cases, although the smaller numbers of this endpoint do not achieve statistical significance. This remains after adjustment for differences in cancer screening habits in vegans;
- We could not demonstrate any significant reduction in risk of breast cancer among vegetarians, although a trend to lower risk (p=0.07) was seen in vegans;
- There appears to be about a 10% non-significant reduction in risk of prostate cancer when comparing raw tomato consumption 4+ times per week to rare consumption; however a similar comparison of intake of cooked/canned tomatoes (lycopene is more available) revealed a significant 28% protective effect;
- Comparing extreme quintiles of isoflavone consumption (half of Adventists eat this at Asian levels) found a 40% significant reduction in risk of breast cancer among the higher users in a calibrated regression;
- Comparing extreme quintiles of dairy consumption, the risk of colon cancer was non-significantly lower in the high dairy group by about 7%. However risk of rectal cancer was a significant 40%/30% lower in the high dairy protein/fat consumption groups. All these results have substantial multivariate adjustment.
Conclusion
The study found that although there was little difference in overall cancer mortality between vegetarians and non-vegetarians, vegetarians were found to have a significantly reduced risk of cancers of the gastrointestinal tract. Additionally, a vegan diet may confer lower risk for overall and female-specific cancer than other dietary patterns.