Few prospective studies have examined cancer incidence among vegetarians. In the Adventist Health Study in California, vegetarians had a significantly lower risk for cancers of the colon and prostate than non-vegetarians, but the risk for breast cancer did not differ significantly between these dietary groups (Fraser, 1999
). In Britain, the Oxford Vegetarian Study suggested no large difference in the incidence of colorectal cancer between vegetarians and non-vegetarians (Sanjoaquin et al, 2004
), whereas the UK Women's Cohort Study suggested that women who did not eat any meat had a lower risk for breast cancer than did meat eaters (Taylor et al, 2007
). The first results from EPIC-Oxford suggested that the incidence of breast cancer did not differ significantly between vegetarians and non-vegetarians (Travis et al, 2008
), that the incidence of colorectal cancer was higher in vegetarians than in meat eaters, that the incidence of lung cancer was lower in fish eaters than in meat eaters, and that the risk for all malignant cancers was lower in fish eaters and possibly lower in vegetarians than in meat eaters (Key et al, 2009
In this paper, we have pooled the individual participant data from the Oxford Vegetarian Study and EPIC-Oxford; hence, this includes data previously reported from these individual studies (Sanjoaquin et al, 2004
; Travis et al, 2008
; Key et al, 2009
). The follow-up time has been extended and, whereas our previous reports included results for only five cancer sites, in this study we have reported the results for 20 cancer sites or groups of sites. The aim of this report is descriptive, and we did not have strong previous hypotheses as to which cancers might show differences in risk between the dietary groups. Therefore, these results should be interpreted cautiously, and for each significant finding we simply give a brief comment in relation to previous evidence and plausibility.
Stomach cancer risk differed significantly between the dietary groups, and was significantly lower in the vegetarians than in the meat eaters, with a similar (non-significantly) low risk among the fish eaters. This observation was based on only 49 cases of stomach cancer. Previous research has suggested that processed meat may increase the risk for stomach cancer, perhaps due to the presence of N
-nitroso compounds (Forman and Burley, 2006
). Therefore, it is plausible that a meat-free diet could be associated with a reduction in the risk for stomach cancer. There is also some evidence that a high intake of fruit and vegetables might reduce the risk for stomach cancer, but the data are not consistent (Forman and Burley, 2006
) and, although on average vegetarians eat more fruit and vegetables than meat eaters, the difference in intake is modest (Key et al, 2009
The risk for cancer of the cervix was significantly higher among vegetarians than among meat eaters, with a similar (non-significantly) high risk among the fish eaters. The principal cause of cervical cancer is human papillomavirus. Dietary factors have been suspected of influencing risk, but no firm conclusions have been drawn (García-Closas et al, 2005
). The increased risks observed in non-meat eaters were based on only 50 cases overall and might be due to non-dietary factors, such as differences in attendance for cervical cancer screening, or to chance.
The risk for ovarian cancer differed significantly between the dietary groups, and was significantly lower among fish eaters than among meat eaters. In a review, Schulz et al (2004)
concluded that high meat consumption may be associated with an increased risk of ovarian cancer. The likely mechanism for such an effect is not clear, and the differences in the risk for ovarian cancer, which we observed, could be due to chance or due to differences in reproductive factors beyond the simple categories of parity and oral contraceptive use for which we were able to adjust.
Prostate cancer risk did not differ significantly between the dietary groups, although there was a significantly lower risk among fish eaters compared with meat eaters. The role of diet in the aetiology of prostate cancer is poorly understood; there is some evidence that high intakes of dairy products might be associated with an increase in risk (Chan et al, 2005
), but to explore this hypothesis further in our data we would need to examine the cancer rates among vegans, among whom there are currently too few cancers to be informative.
The risk for bladder cancer was lower among vegetarians than among meat eaters, based on 85 cancers overall. Some previous studies have suggested that certain meats, such as bacon, might increase the risk for bladder cancer, perhaps due to preformed nitrosamines (Lijinsky, 1999
; Michaud et al, 2006
), and this area deserves further investigation.
We observed a striking difference between the dietary groups in the risk for the group of cancers of the lymphatic and haematopoietic tissues, on the basis of 257 cancers overall. The risk for these cancers was not significantly reduced among fish eaters, but among vegetarians the risk was substantially lower than that among meat eaters. Among the three major cancer types contributing to this grouping, the risks for non-Hodgkin's lymphoma and multiple myeloma, but not leukaemia, were significantly lower in vegetarians than in meat eaters. Previous research has suggested inconsistently that consumption of meat and/or exposure to live animals and raw meat among farmers and butchers might be associated with an increased risk for some of these cancers (Zhang et al, 1999
; Alexander et al, 2007
). Potential mechanisms could include mutagenic compounds and viruses (Cross and Lim, 2006
; Alexander et al, 2007
We did not observe any significant difference in the incidence of colorectal cancer between the dietary groups. Our earlier publications from the Oxford Vegetarian Study and EPIC-Oxford also did not report a reduction in risk for colorectal cancer among vegetarians (Sanjoaquin et al, 2004
; Key et al, 2009
). We also noted previously in EPIC-Oxford, that the incidence of colorectal cancer among vegetarians was identical to that in the general population of England and Wales (standardised incidence ratio 102% (95% CI: 80–129); Key et al, 2009
). In the Adventist Health Study, a lower risk for colon cancer was observed among vegetarians compared with non-vegetarians (rectal cancer was not reported; Fraser, 1999
). In our pooled analysis of mortality in five prospective studies, comprising the Adventist Mortality Study, the Adventist Health Study, the Health Food Shoppers Study, the Oxford Vegetarian Study and the Heidelberg Study, we observed no difference between vegetarians and non-vegetarians in mortality from colorectal cancer (Key et al, 1999
). The 2007 report from the WCRF/AICR concluded that the evidence that high intakes of red and processed meat cause colorectal cancer is convincing (WCRF/AICR, 2007
). In the largest single prospective study on this relationship, Cross et al (2007)
reported that the risk for colorectal cancer was increased by 20% at moderate red meat intakes (equivalent to ~86
g per day in men and ~44
g per day in women). Meat intake among meat eaters in EPIC-Oxford was estimated as 78.1 and 69.7
g per day in men and women, respectively (Key et al, 2009
), lower than intakes reported in the National Diet and Nutrition Survey for the United Kingdom, but still providing a substantial difference in intake between meat eaters and non-meat eaters. It is possible that this study did not have enough power to detect a moderate reduction in the risk for colorectal cancer among vegetarians, but our null findings on vegetarians suggest that the relationship of meat with the risk for colorectal cancer requires further research.
Total cancer incidence was significantly lower among both fish eaters and vegetarians than among meat eaters. This difference in total cancer incidence between meat eaters and non-meat eaters could not be ascribed to any one of the major cancer sites examined. We are unaware of other data comparing total cancer incidence in meat eaters and non-meat eaters, and the reason for this small difference is not known. More data are needed to further our understanding of this observation, which if confirmed is likely to be due to differences for specific cancer sites.
The results presented in this study are simply descriptive of the incidence of cancer in fish eaters and vegetarians relative to meat eaters. More detailed analyses of individual cancer sites are needed to explore, for example, whether the differences observed might be linked to particular types of meat or to other dietary or lifestyle characteristics of non-meat eaters that were not adjusted for in the current analysis.
A potential weakness of this type of study is the accuracy of the assessment of vegetarian status. The diet group was assigned on the basis of the answer to four questions, asking specifically about whether participants ever ate meat, fish, dairy products and eggs. When the diet group in EPIC-Oxford was assigned on the basis of answers to the same four questions in a follow-up questionnaire 5 years later, 85% of the vegetarians were allocated to the same diet group as at the time of recruitment (Key et al, 2009
), suggesting that the assessment of vegetarian status is accurate and stable over at least several years, and may be a substantially more stable dietary characteristic than epidemiological estimates of nutrient intakes.
In conclusion, this study suggests that the incidence of all malignant neoplasms combined may be lower among both fish eaters and vegetarians than among meat eaters. The most striking finding was the relatively low risk for cancers of the lymphatic and haematopoietic tissues among vegetarians.