Fruit and vegetable consumption in adolescence and early adulthood and risk of breast cancer: population based cohort study
BMJ 2016; 353 doi: https://doi.org/10.1136/bmj.i2343 (Published 11 May 2016) Cite this as: BMJ 2016;353:i2343
All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
Dear Editor:
The recent article by Farvid et al.,1 raises two important questions. First, when will the nutrition community acknowledge the empirical refutation of memory-based dietary assessment methods (M-BMs; e.g., questionnaires, interviews, surveys)? 2-4 Second, when does the persistent refusal to acknowledge the empirical refutation of one’s method become scientific misconduct?
My recent work 2 3 clearly demonstrates that M-BMs are pseudoscientific, produce data that are physiologically implausible 4 and incompatible with life. Therefore, these data are inadmissible as scientific evidence. 2 3 It should be apparent that dietary data that are not compatible with life cannot be valid. Or stated more simply, the diets reported in studies such as the Nurses’ Health Study II could not support human life if consumed on a daily basis and therefore are inadmissible as scientific evidence. The conclusions of my work are supported by many decades of research 5 and suggest that M-BMs collect little more than uncorroborated anecdotal (i.e., qualitative) estimates of food and beverage consumption. 2 3 Given the voluminous evidence and many decades of empirical refutation, at what point does it become scientific misconduct to continue to publish physiologically implausible dietary data?
References
1. Farvid MS, Chen WY, Michels KB, et al. Fruit and vegetable consumption in adolescence and early adulthood and risk of breast cancer: population based cohort study. BMJ 2016;353.
2. Archer E, Pavela G, Lavie CJ. The Inadmissibility of What We Eat in America and NHANES Dietary Data in Nutrition and Obesity Research and the Scientific Formulation of National Dietary Guidelines. Mayo Clin Proc 2015;90(7):911-26.
3. Archer E, Pavela G, Lavie CJ. A Discussion of the Refutation of Memory-Based Dietary Assessment Methods (M-BMs): The Rhetorical Defense of Pseudoscientific and Inadmissible Evidence. Mayo Clinic Proceedings 2015;90(12):1736-38.
4. Archer E, Hand GA, Blair SN. Validity of U.S. nutritional surveillance:National Health and Nutrition Examination Survey caloric energy intake data, 1971-2010. PLoS One 2013;8(10):e76632.
5. Goldberg GR, Black AE, Jebb SA, et al. Critical evaluation of energy intake data using fundamental principles of energy physiology: 1. Derivation of cut-off limits to identify under-recording. Eur J Clin Nutr 1991;45(12):569-81.
Competing interests: No competing interests
While there is nothing implausible about the hypothesis behind this study, that the risk of cancer is modified early in life by micronutrients and other factors, putatively supplied in this study by fruit and vegetables, although potentially also able to be sourced from unprocessed legumes, nuts and seeds, and high-quality animal foods, it is unfortunate that the data collection method used, the nostalgic recreation of the childhood diet by adults, really amounts to inviting respondents to invent data.
For example, I can vaguely remember the vegetables I ate as an adolescent because they were part of familiar meals my mother cooked, though how often she cooked them is now unclear, but I have no accurate recall of my opportunistic consumption of fruit. Yet the exclusion criteria ruled out data from respondents who failed to complete the fruit and vegetable sections of the questionnaire, most likely from motives of honesty rather than negligence, possibly exaggerating any correlation.
In table 3, the association between adolescent fruit and cancer incidence lacks a dose response. All we can really say is that there is an increased risk from very low fruit consumption. However, this is confounded by the higher rate of smoking in this group. In table 4, for vegetables, there is again a lack of dose response. (the ORs for all breast cancer cases in table 5, fruit and veg combined, run 1, 0..85, 0.79, 0.82, 0.82 from lowest to highest intakes in the fully adjusted model).
If we take the data at face value – which we really should be reluctant to do, despite the inherent plausibility of the hypothesis – it does not support the value of a “plant-based diet” in cancer prevention, as opposed to the common-sense approach of including some plant wholefoods in the diet for nutritional reasons.
The main harm that results from such prestigious researchers publishing such low-quality data, is that the higher-quality data that does exist from other sources tends to become submerged in a swamp of biased results known as “the totality of the evidence”. For example, a much better quality data set, the Malmo Diet and Cancer study, which used a food frequency questionnaire with a 7-day food diary and an interview to collate data and did not use retrospective diet recall, did not support the link between fruit and vegetables and breast cancer (ref 10 in this paper) but did reveal other associations; a protective association with fibre and folate, and (in the case of post-menopausal breast cancer) an association with a higher intake of omega 6 polyunsaturated fat, and with fat made from vegetable oil, in 2 case-control analyses broadly consistent with a prospective cohort all-cancer mortality analysis.[1]
“Saturated fat and the omega3-omega6 fatty acid ratio were not related to increased risks, but positive trends were seen for total (p = 0.031), monounsaturated (p = 0.002), and polyunsaturated fat (p = 0.0009), especially omega6 fatty acids and the polyunsaturated-saturated fat ratio (p = 0.004). With mutual adjustment for different types of fat, an elevated risk remained significant in the highest omega6 fatty acid quintile (RR= 2.08, 95% CI 1.08-4.01).”[2]
“Fat from fermented milk products was negatively associated with breast cancer risk (trend, P = 0.003). The highest quartiles of vegetable oil-based dietary fats (odds ratio, OR = 1.74; confidence interval, CI = 1.12-2.72) and dried soup powders (OR = 1.59; CI = 1.04-2.43) showed positive associations.”[3]
The latter result may well in fact be consistent with the Harvard hypothesis, if the foods associated with breast cancer in Malmö, vegetable-oil based fats and powdered soups, are markers of a micronutrient-poor lifestyle. Perhaps it all comes down to good nutrition and metabolic health in the end, and perhaps the effect of the inessential dietary components, the ones marketed in superfoods and as phytochemical supplements, is relatively minor at a population level. This should at least be the null hypothesis in such diet-health calculations.
[1] Leosdottir M, Nilsson PM, Nilsson JA, Månsson H, Berglund G. Dietary fat intake and early mortality patterns--data from The Malmö Diet and Cancer Study. J Intern Med. 2005 Aug;258(2):153-65.
[2] Wirfält E, Mattisson I, Gullberg B, Johansson U, Olsson H, Berglund G. Postmenopausal breast cancer is associated with high intakes of omega6 fatty acids (Sweden). Cancer Causes Control. 2002 Dec;13(10):883-93.
[3] Wirfält E, Mattisson I, Gullberg B, Olsson H, Berglund G. Fat from different foods show diverging relations with breast cancer risk in postmenopausal women. Nutr Cancer. 2005;53(2):135-43.
Competing interests: No competing interests
Dear Editor
Kudos to Marym S Farvid et al (1) for the article that can potentially generate interest from researchers and laypeople alike.
The article recollects painstakingly 24 hour recall thrice for generating dietary history dating back to year, which creates a unique scenario. Besides taking the frequency of food and dietary intake, the consumption weight (gm) of the daily fruit intake could have favoured the study as every nutrient e.g vitamin c, alpha carotene and beta carotene depends on the total amount of fruits or vegetable intake rather than their frequency.
Another point that comes to my mind is why receptor status of HER 2 NEU is not taken against the risk of breast cancer. This is because according to some studies HER 2 NEU expressing breast carcinomas have increased COX - 2 expression(2) and Beta carotene is associated with decreased expression of COX-2 and increased apoptosis in various cancer cell lines(3).
Nevertheless this study has added to the current literature, potentially opening gates of further research that may include randomised/non randomised trials or observational studies involving intake of fruit and vegetables alongwith established protocols breast cancer treatment.
References
1. Farvid Maryam S, Chen Wendy Y, Michels Karin B, Cho Eunyoung, Willett Walter C, Eliassen A Heather et al. Fruit and vegetable consumption in adolescence and early adulthood and risk of breast cancer: population based cohort study BMJ 2016; 353 :i2343.
2. Subbaramaiah K, Norton L, Gerald W, Dannenberg AJ. Cyclooxygenase-2 is Overexpressed in HER-2/neu-Positive Breast Cancer. Evidence for Involvement of AP-1 and PEA3.
3. Developmental windows of breast cancer risk provide opportunities for targeted chemoprevention Holly A. Martinson, Traci R. Lyons, Erin D. Giles, Virginia F. Borges, Pepper Schedin. Exp Cell Res. 2013 July 1; 319(11): 1671–1678.
Competing interests: No competing interests
Farvid et al conclude in their article published on 11 May 2016 that there could be an association between high fruit and vegetable intake, especially in adolescence, and a lower risk of breast cancer. It is clear that their study design could not show causality but a mere association. This is a very important point to note, especially in the face of certain recent, widely publicised studies which have been misinterpreted and fed to the layperson, and credit goes to the authors for acknowledging this.
Furthermore, the authors tried to minimise the effect of confounding factors that could have been associated with higher fruit and vegetable intake (for example BMI, smoking status, parity) by adjusting for them in their analysis. Unfortunately, the nature of their data source (Nurses' Health Study II questionnaire) meant that they could not and did not adjust for all other potential confounding factors (for example lower socioeconomic status, level of education and awareness of breast cancer, breastfeeding).
Certainly, their hypothesis fulfils the criterion of biological plausibility and it can be argued that it can only be good for us all to adopt a diet higher in fruit and vegetables for other conditions with a proven link. However, the points made above, in addition to the uncertainty added by the self-reported nature of diet makes this study difficult to convincingly defend.
Competing interests: No competing interests
Dear Editors
Mr Thompson in his rapid response (http://www.bmj.com/content/353/bmj.i2343/rr-1) kindly pointed out one significant flaw with regard to this paper, but this is only one of a few.
In addition to difficulties associated with recollection of dietary intake at least 15 years apart, the authors cannot adequately account for various problems associated with dietary intake estimation including (but not exclusively):
Problem with portion estimation by individuals
Association of access to fresh fruit/vegetable with socioeconomic status prior to or post nurse qualification
Changes in working status and related socioeconomic status since initial recruitment not accounted for.
Which is why the applicability of such data analysis is dubious from the start, which can be construed as a futile attempt at data dredging initiated 20 years ago by the supplementary questionnaire.
Certainly I doubt the Nurses’ Health Study II needed this research to ruin its reputation
Competing interests: No competing interests
The adolescent fruit consumption data used in this study appears to be based on a dietary recall questionnaire completed in 1998 by subjects who were already aged at least 27 in 1991 when they completed the "early adulthood" questionnaire.
So the youngest participant providing dietary recall data of their adolescent diet was aged 34.
Really ?
I'm sorry, but this sort of vague epidemiology based on distant historical recall really isn't worthy of our attention.
Competing interests: No competing interests
This important piece of research is well defined and carefully narrow limited.
I believe that it would be teaching to remind, that in their influential "Causes of cancer" article (PMID: 7017215 ) and book (ISBN-10: 0192613596) R. Doll and R. Peto attracted the world's attention to the association of the breast cancer with diet fat. It was so strong and compelling, that together with cardiovascular research of the time, it makes the animal fat an enemy of the humanity for at least 20 years.
Now the connection of the breast cancer to diet fat appears so negligible, that in this text it is mentioned only as a confounder to be adjusted.
Alongside with the main result of the study, the association of the low breast cancer risk with increased consumption of carotenoids is presented. Again, it is given as a positive finding without mentioning the very negative results of the controlled trials of the carotenoids in cancer prevention. Without the proper context it may feed the worm of vitamin industry and turn the research efforts to another desert, similar to vitamin D Sahara (of last dozen years), where so many scientists and so many patients lost their time and public money.
Competing interests: No competing interests
Prior to International Nurses Day, Farvid and colleagues brought a scientific gift for nurses and women globally, i.e., a timely scientific article with solid evidence [1]. Based on 22 years’ follow-up amongst female registered nurses, Farvid and colleagues successfully addressed the underlying clues between higher fruit intake during adolescence and lower risk of breast cancer. The mutational theory for the pathogenesis of breast cancer identifies driver mutations, most of which arise in somatic cells during the lifetime. Whole genome sequencing of 560 cases with breast cancer by Nik-Zainal and colleagues reveals the landscape of somatic mutations [2].
Farvid and colleagues analyzed dietary impact on the risk of breast cancer among a large cohort; whereas Nik-Zainal and colleagues studied the driver mutations and novel cancer genes using next-generation sequencing. In combining both articles, it might be reasonable to propose that higher fruit intake during adolescence might protect somatic cells from detrimental driver mutations. Such additional linkage analysis might be critical important for the etiology of breast cancer, the welfare of patients and the public.
It has been well established that fruit and vegetables are linked with increased bone mass and decreased urinary calcium excretion, owing to abundant amounts of nutrients, such as vitamin C and K, phytochemicals [3 4]. Moreover, diets high in fruits and vegetables have a lower acid load and consequently uphold a positive calcium balance [5]. Accumulating evidence indicates the protective effect of fruits and vegetables against osteoporosis among children, adults and elders [6-9].
In fact, we released a science popular essay through our WeChat platform as "Spine Truth" (jzglyl) on Mar 6, 2016 (Figure 1).
Collectively, higher fruit and vegetable intakes are beneficial for the public in terms of lower risk of cancer, osteoporosis and healthy bone mass.
References
1. Farvid MS, Chen WY, Michels KB, et al. Fruit and vegetable consumption in adolescence and early adulthood and risk of breast cancer: population based cohort study. BMJ 2016;353:i2343 doi: dx.doi.org/10.1136/bmj.i2343 [published Online First: Epub Date]|.
2. Nik-Zainal S, Davies H, Staaf J, et al. Landscape of somatic mutations in 560 breast cancer whole-genome sequences. Nature 2016 doi: 10.1038/nature17676[published Online First: Epub Date]|.
3. Tucker KL, Hannan MT, Chen H, et al. Potassium, magnesium, and fruit and vegetable intakes are associated with greater bone mineral density in elderly men and women. The American journal of clinical nutrition 1999;69(4):727-36
4. Trzeciakiewicz A, Habauzit V, Horcajada MN. When nutrition interacts with osteoblast function: molecular mechanisms of polyphenols. Nutrition research reviews 2009;22(1):68-81 doi: 10.1017/S095442240926402X[published Online First: Epub Date]|.
5. New SA. Intake of fruit and vegetables: implications for bone health. The Proceedings of the Nutrition Society 2003;62(4):889-99 doi: 10.1079/PNS2003310[published Online First: Epub Date]|.
6. New SA, Robins SP, Campbell MK, et al. Dietary influences on bone mass and bone metabolism: further evidence of a positive link between fruit and vegetable consumption and bone health? The American journal of clinical nutrition 2000;71(1):142-51
7. Prynne CJ, Mishra GD, O'Connell MA, et al. Fruit and vegetable intakes and bone mineral status: a cross sectional study in 5 age and sex cohorts. The American journal of clinical nutrition 2006;83(6):1420-8
8. Liu ZM, Leung J, Wong SY, et al. Greater fruit intake was associated with better bone mineral status among Chinese elderly men and women: results of Hong Kong Mr. Os and Ms. Os studies. Journal of the American Medical Directors Association 2015;16(4):309-15 doi: 10.1016/j.jamda.2014.11.001[published Online First: Epub Date]|.
9. Xie HL, Wu BH, Xue WQ, et al. Greater intake of fruit and vegetables is associated with a lower risk of osteoporotic hip fractures in elderly Chinese: a 1:1 matched case-control study. Osteoporosis international : a journal established as result of cooperation between the European Foundation for Osteoporosis and the National Osteoporosis Foundation of the USA 2013;24(11):2827-36 doi: 10.1007/s00198-013-2383-9[published Online First: Epub Date]|.
Competing interests: Supported by National Natural Science Foundation of China (No. 81270028 and 81572182).
Re: Fruit and vegetable consumption in adolescence and early adulthood and risk of breast cancer: population based cohort study
We appreciate the comments raised by BMJ readers on our study. As we acknowledged in our article, recall of adolescent diet as an adult will be affected by measurement error, because it relies primarily on memory of diet many years earlier. We were not able to conduct the ideal validation study of the high school dietary questionnaire, which would have involved collecting detailed weighed dietary records and biomarkers from high school students as “gold standards” and having participants fill out the retrospective questionnaire several decades later. However, no such data exist, so we used several methods to assess validity somewhat indirectly, and these all suggest that participants can provide useful information although probably with a lesser degree of validity than when reporting current diet (1, 2). These indirect assessments probably underestimate true validity because the comparison methods themselves have error, so we need to interpret these correlations somewhat qualitatively. Importantly, in our analysis we prospectively examined risk of breast cancer after we collected the information on adolescent diet, thus eliminating bias related to the diagnosis of breast cancer. Notably, the associations we found with high school diet are stronger than we see for adult diet, despite the longer recall, which is consistent with much prior evidence that breast tissue is more sensitive to carcinogenic influences earlier in life. Ideally, the findings related to adolescent intake of fruits and vegetables and risk of breast cancer would be evaluated in a randomized trial, but this is clearly not feasible.
Furthermore, we have ample evidence that our questionnaire does measure adult fruits and vegetables reasonably well, based on correlations with plasma carotenoids (3) and prediction of other major outcomes for which benefits of fruits and vegetables are strongly supported (4, 5). Validity for assessing other aspects of diet, including use of objective biomarkers, has been extensively documented elsewhere (6). We would also note that the comments by Archer were directed at all self-reported information, not specifically our dietary questionnaire. His campaign to discredit any self-reported information, funded by Coca-Cola (7), is based on an argument that energy intake calculated from dietary assessments does not predict long term weight gain, an unrealistic criterion for both objective and self-reporting methods, because precision of less than 1% for both dietary intake and physical activity would be needed (8). Meanwhile, based on data, much of it derived from our questionnaire, showing clear relationships between soda consumption and risks of diabetes and weight gain, soda consumption in the US has dropped by about 25%, the increase in childhood obesity has been arrested and may be declining (9), and as recently announced by the CDC the incidence of type 2 diabetes in the US has declined by about 20% in the last several years (10). (In addition to the reduction in soda consumption, the near elimination of trans fat, again largely based on data using our questionnaire, has almost certainly also contributed to the decline in type 2 diabetes). Thus, we have been able to show that diet assessed by our questionnaire has good validity compared with detailed assessment methods and biomarkers, predicts major disease outcomes, can influence dietary recommendations, can lead to major national changes in diet, and can reverse major epidemics.
Although the prevalence of smoking was higher among women with low intake of adolescent fruit consumption, results were essentially unchanged after adjusting for smoking and most known risk factors for breast cancer (Model 2). Furthermore, no significant interaction was observed between smoking status and either adolescent or adulthood fruit and vegetable intake. In response to another comment, our study does not have sufficient power to evaluate the association between fruit and vegetable intake and risk of molecular subtype of breast cancer. However, it needs to be evaluated in future studies.
References
1. Maruti SS, Feskanich D, Rockett HR, Colditz GA, Sampson LA, Willett WC. Validation of adolescent diet recalled by adults. Epidemiology. 2006;17:226-9.
2. Maruti SS, Feskanich D, Colditz GA, Frazier AL, Sampson LA, Michels KB, Hunter DJ, Spiegelman D, Willett WC. Adult recall of adolescent diet: reproducibility and comparison with maternal reporting. Am J Epidemiol. 2005; 161:89-97.
3. Hendrickson SJ, Willett WC, Rosner BA, Eliassen AH. Food predictors of plasma carotenoids. Nutrients. 2013;5:4051-66.
4. Muraki I, Imamura F, Manson JE, Hu FB, Willett WC, van Dam RM, Sun Q. Fruit consumption and risk of type 2 diabetes: results from three prospective longitudinal cohort studies. BMJ. 2013;347:f5001.
5. Joshipura KJ, Hu FB, Manson JE, Stampfer MJ, Rimm EB, Speizer FE, Colditz G, Ascherio A, Rosner B, Spiegelman D, Willett WC. The effect of fruit and vegetable intake on risk for coronary heart disease. Ann Intern Med. 2001;134:1106-14.
6. Willett WC, Lenart E. Reproducibility and validity of food frequency questionnaires. In: Nutritional Epidemiology. pp. 96-141. Oxford University Press; 2013.
7. http://well.blogs.nytimes.com/2015/08/09/coca-cola-funds-scientists-who-...
8. Subar AF, Freedman LS, Tooze JA, Kirkpatrick SI, Boushey C, Neuhouser ML, Thompson FE, Potischman N, Guenther PM, Tarasuk V, Reedy J, Krebs-Smith SM. Addressing Current Criticism Regarding the Value of Self-Report Dietary Data. J Nutr. 2015;145:2639-45.
9. Ogden CL, Carroll MD, Fryar CD, Flegal KM. Prevalence of obesity among adults and youth: United States, 2011–2014. NCHS Data Brief, No. 219, November 2015. Available at: http://www.cdc.gov/nchs/data/databriefs/db219.pdf
10. http://www.cdc.gov/diabetes/statistics/incidence/fig2.htm
Competing interests: No competing interests