Epidemiology

General supplement use, subsequent use and cancer risk in the UK Women’s Cohort Study

Abstract

Background/Objectives:

To determine whether general dietary supplement use is associated with cancer risk in UK women and to estimate risks related to use at one and two recording points.

Subjects/Methods:

Cox’s proportional hazard regression models were used to estimate cancer risks for 32 665 middle-aged women in the UK Women’s Cohort Study relating to any current supplement use recorded in a baseline questionnaire. During a median follow-up of 15 years, there were 3936 registered cancer incidences, including 1344 breast, 429 smoking-related and 362 colorectal cancers. Cancer risks for 12 948 of these women, who also completed questionnaires on average 4.4 years later, were estimated in relation to any supplement use at both time points (1527 cancers, including 561 breast, 131 smoking-related and 141 colorectal cancers). Adjustments were made for baseline confounders.

Results:

Total smoking-related cancers were associated with baseline supplement use (hazard ratio (HR)=1.41, 95% confidence interval (CI): 1.10, 1.81) compared with non-use, but not associated with use at both recording points (HR=1.29; 95% CI: 0.78, 2.13) compared with use at neither. There was no evidence of the associations between total, colorectal or breast cancers and baseline supplement use, or use at both recording points. In sub-analyses, no significant associations with breast cancer were found for premenopausal or postmenopausal baseline users, or similarly for use at both points (HR=1.35, 95% CI: 0.91, 2.01; and HR=0.93, 95% CI: 0.68, 1.26, respectively).

Conclusions:

There was evidence that general supplement use was associated with increased smoking-related cancer risk, but there was no evidence of associations with total, colorectal and breast cancers.

Introduction

A large proportion of UK women take supplements; 41% of women in the 2008/9 National Diet and Nutrition Survey reported using supplements in the previous year.1 Some women may take supplements to reduce their risk for chronic diseases.2 However, the 2007 World Cancer Research Fund review clearly states that supplements are not recommended for cancer prevention, and reports no convincing evidence from their systematic reviews that specific micronutrients in supplement form protect against the risk of developing cancers that affect women.3 Furthermore, increased lung cancer risks were found for smokers who used high-dose β-carotene supplements.4,5

Supplementation with a variety of micronutrients may not be protective either; a meta-analysis of Randomised Control Trials (RCTs) showed no association between multivitamin supplement use and cancer mortality.6 Nevertheless, multivitamin use has recently been reported to reduce total cancer risk in men in a large US RCT;7 men, though, tend to have lower baseline antioxidant status than women indicative of a poorer diet.8 Two large US cohort studies reported no association between multivitamin supplement use in women and total cancer incidence, or cancer at the major sites such as breast, colorectal and lung.9,10 Ten-year, long-term use of multivitamins, which contain low-dose micronutrients, was also not associated with lung cancer in the Vitamins and Lifestyle (VITAL) cohort study.11 High-dose users tend to take a number of supplement types,12 and therefore are likely to supplement with a range of micronutrients. However, the majority of previous research has analysed risks relating to single supplements rather than multiple or general supplement use.

Although RCTs are less prone to bias than observational studies, supplementation over the intervention period may be insufficient to affect long-term risks, and these periods are usually substantially shorter than follow-up periods in most prospective observational studies. In addition, unlike drug trials, members of control groups can easily obtain supplements. On the other hand, it is possible that the sporadic nature of supplement use in free-living populations, as highlighted in European Prospective Investigation into Cancer and Nutrition (EPIC)-Heidelberg,13 might explain the lack of associations or inconsistencies in results in cohort studies. However, only the analyses of the EPIC-Heidelberg cohort and a Japanese cohort study have assessed whether general supplement use at more than one recording point was associated with cancer incidence,14 or with cancer mortality,15 and no associations were found for users who took supplements at more than one recording point. However, there was evidence from these cohorts that users at more than one recording point had a healthier diet compared with past and new users as well as non-users.13,14

This study prospectively analyses the relationship between current use of any supplement type and total cancer, smoking-related cancer,16,17 colorectal cancer and breast cancer risk in UK women. In secondary analyses, the cancer risks for women using supplements at two recording points, and the risks for use at only one of these points, are compared with the risks for women not taking supplements at either point. In addition, the characteristics of UK women in these different supplement user groups are compared.

Materials and methods

Study population

UK Women’s Cohort Study (UKWCS) recruitment data were gathered between 1995 and 1998 from 35 367 women aged between 33 and 74 years who completed a 217-item validated food frequency questionnaire.18, 19, 20 This national cohort of mainly Caucasian, well-educated, middle-class, middle-aged, married women was designed to compare disease incidence in vegetarians, fish eaters and meat eaters.18 At recruitment, 34 958 (99%) provided information about whether or not they took supplements.

Women with any prevalent malignant cancers recorded in UK cancer registries before baseline cohort entry were excluded. This provided 32 665 women for the risk analysis, which compared women who took supplements at baseline with those who did not. Over the median follow-up period of 15 years, there were 3936 malignant cancer incidences (including non-melanoma skin cancers) ascertained from UK cancer registries via the UK Office of National statistics. These included 1344 breast and 362 colorectal cancers diagnosed by the censor date 1 October 2011. There were also 429 smoking-related cancers16,17 (lung (172), cervical (27), bladder (53), kidney (48), oesophageal (47), stomach (24) and pancreas (66)).

To explore the stability of general supplement use between two survey points, further secondary analyses were undertaken for 12 948 of the above women who also completed supplement questions on the UKWCS follow-up survey between 2 and 5 years after baseline (4.4 years on average). These women were split into three categories. ’Users at both’ were defined in these analyses as women who were taking any type of supplement at both the baseline and the second survey. ‘New/past users’ reported taking supplements at only one of the two survey time points and ‘never-users’ reported no supplement use at either. Over a median follow-up period of 15 years from baseline, there were 1527 incident cancers in total and 561 incident breast, 131 smoking-related and 141 colorectal cancers registered to the censor date 1 October 2011 in these 12 948 women. One hundred and forty eight (26%) of the 561 breast cancer cases and 308 (20%) of the 1527 total cancer cases occurred between baseline and the second survey in the UKWCS.

Data collection

General supplement use at baseline was determined by means of a questionnaire using self-reported yes/no answers to:

Do you take any vitamins, minerals, fish oils or other food supplements?

General supplement use at the second survey was determined by means of a questionnaire using self-reported yes/no answers to:

Do you presently use any dietary supplements?

Supplements were defined on this second survey questionnaire as vitamins, minerals, fibre, fish oils or other food supplements. In addition, if participants did not answer yes to the above questions but provided details of any type of supplements taken, regardless of amount taken, these women were designated as being general supplement users. The most popular supplement types taken daily were determined at this second survey where women were asked to indicate the frequency of types they took from a list of 16 types provided on the questionnaire.

Covariates were derived from the health and lifestyle part of the baseline questionnaire, except for total alcohol intake and total energy intake which were derived from the baseline food frequency questionnaire.

Statistical analyses

Characteristics of women in the different supplement use categories were described using means and percentages. Significant differences between means of baseline users and non-users were established using t-tests and significant differences between categories were established using χ2 test. Any significant trends across groups from ‘never-users’ to ‘new/past users’ to ‘users at both’ were determined using linear regression followed by tests for linear hypotheses for means, or using χ2 test for trend for dichotomous variable percentages.

Cox proportional hazard regression was used to provide hazard ratios (HRs) to estimate cancer risks in relation to supplement use for women in the UKWCS. The reference groups in the main analyses were ‘non-users’ at baseline, whereas in the secondary analyses cancer risks for ‘users at both’ and ‘new/past users’ categories were compared with ‘never-users’. Probability weighting was used to produce estimated average risks representative of the UK population of women, needed because of the recruitment of substantially higher proportions of vegetarians and fish eaters into the cohort compared with the UK population. Vegetarians and fish eaters were weighted by 0.27 and 0.43, respectively. Adjustments were made in the total cancer, smoking-related cancer and colorectal cancer analyses for age, body mass index (underweight, normal, overweight, obese), education (none, up to degree level, degree), smoking status (never, past, current occasionally, current every day), minutes of sweating exercise per week, alcohol intake (g/day), total energy intake (kcal/d) and dietary type (mainly meat, oily fish, other fish eater, vegetarian as described elsewhere18). In addition to the above, the breast cancer analyses were adjusted for contraceptive pill use (never, past, current), hormone replacement therapy (HRT) (never, past, current), age at menarche, and parity, but not smoking status. Although there was no interaction between supplement use at baseline and menopausal status in relation to breast cancer risks in likelihood ratio tests for interactions, HRs were examined by menopausal status, as commonly practised in breast cancer risk analyses. Menopausal status at baseline was determined from responses to a number of questions relating to last natural period, HRT and contraceptive pill use, hysterectomy and ovary removal.21 Sensitivity analyses were undertaken to make additional adjustments for family history of cancer in first-degree relatives (missing data was >5%) and in the breast cancer analysis for estimated weeks of breastfeeding. Other sensitivity analyses excluded incident cancers within 2 years of baseline. Further sensitivity analysis, with a median follow-up time of 10.6 years, was undertaken relating to use at two time points where the time-to-event calculation was started from the second survey date instead of from baseline, thereby excluding incident cases diagnosed between the surveys. Analyses were carried out using Stata version 12 (Timberlake Consultants UK, London, UK) and results were based on a significance level of P<0.05.

Results

Of the 32 665 eligible women, 62% were supplement users at baseline and 54% of the 12 948 women at the second survey had taken supplements at both recording points, and were classed as users at both. At the second survey, 25% were new/past users (had taken supplements at either recording point but not at both) and 21% were never-users. Of the 8915 (69%) women who were currently taking supplements at the second survey, 7010 (79%) had also taken supplements at baseline and were classed as users at both. On the second survey questionnaire, 27% of all women reported taking fish oils, 25% reported taking multivitamins/minerals, 19% were taking primrose/starflower oil, 18% reported taking vitamin C and 14% were taking calcium supplements on a daily basis. These were the most popularly named supplement types, and for each of these there were statistically significant differences in intake between menopausal statuses: multivitamins and primrose/starflower oil were more likely to be used by premenopausal women, whereas cod liver oil and the others were more likely to be taken by women of postmenopausal age.

There were statistically significant differences between general supplement users and non-users at baseline for the majority of characteristics listed in Table 1. In particular, users reported a higher fruit and vegetable intake, a lower meat intake, lower alcohol intake and reported doing more vigorous exercise compared with non-users. Users were also more likely to have a family history of any cancer. However, there were no significant differences in level of education.

Table 1 Characteristics of all supplement users and non-users at baseline in the UK Women’s Cohort Study

Similarly as observed in Table 2, there were increasingly healthier behaviours relating to exercise, alcohol use, fruit and vegetable intake and meat intake from never-users (at baseline and second survey), through new/past users, to users at both recording points. There was a decreasing trend for HRT use, number of children and estimated cumulative breastfeeding, but an increase in trend for supplement use for ex-smokers. However, there were no significant differences between user frequency in relation to family history of cancer.

Table 2 Baseline characteristics of women in the UK Women’s Cohort Study according to supplement use at baseline and second survey for never-users, past/new users, users at both

As observed in Table 3 in the analysis of supplement use assessed at baseline for all women, there was no statistically significant difference in total cancer, colorectal cancer or breast cancer risk between all supplement users and non-users either in the unadjusted or adjusted analyses (adjusted HR=1.06, 95% confidence interval (CI): 0.98, 1.14; HR=1.13, 95% CI: 0.87, 1.46; and HR=1.01, 95% CI: 0.89, 1.15, respectively). There was no evidence of interactions between supplement use and menopausal status (P=0.13) or HRT use on breast cancer risk. Furthermore, in the sub-analysis by menopausal status, HRs were not statistically significant (adjusted HR=0.96, 95% CI: 0.81, 1.13; and HR=1.07, 95% CI: 0.88, 1.30 for postmenopausal women and premenopausal, respectively, as shown in Table 3). However, an association was found for smoking-related cancers (HR=1.41, 95% CI: 1.10, 1.81; P=0.007), which remained after excluding cancers occurring within 2 years of baseline (HR=1.36, 95% CI: 1.05, 1.76; P=0.02).

Table 3 Cancer risks for all supplement users at baseline in the UK Women’s Cohort Study compared with non-users (N=32 665)

Similarly in the secondary analysis exploring consistency of supplement-taking between baseline and the second survey, compared with never-users there was no significant difference in the risk for total cancer, colorectal cancer and breast cancer in the adjusted analyses for new/past users (HR=0.96, 95% CI: 0.81, 1.14; HR=0.71, 95% CI: 0.40, 1.26; and HR=0.98, 95% CI: 0.73, 1.30, respectively) and users at both recording points (HR=1.01; 95% CI: 0.88, 1.17; HR=1.00, 95% CI: 0.63, 1.57; and HR=1.08; 95% CI: 0.85, 1.38, respectively) as observed in Table 4. Although the point estimates were raised in the premenopausal sub-analysis they were not statistically significant (HR=1.17, 95% CI: 0.74, 1.84; and HR=1.35, 95% CI: 0.91, 2.01, respectively, for past/new users and users at both recording points). In addition, no evidence of association was found for smoking-related cancers for past/new users (HR=0.97 95% CI: 0.53, 1.79) or for users at both recording points (HR=1.29; 95% CI: 0.78, 2.13). Additional adjustment for estimated cumulative breastfeeding duration in the breast cancer analyses and for family history of cancer had little effect on HRs and confidence intervals in the above analyses.

Table 4 Cancer risks for users at both recording points, and for past/new users compared with never-users, according to any supplement use at baseline and second survey in the UK Women’s Cohort Study (N=12 948)

The results of the sensitivity analysis, which excluded cases diagnosed within 2 years of baseline or diagnosed between supplement-taking surveys, did not affect the overall conclusion of this research (data not shown).

Discussion

There was little evidence of associations between general supplement use and total cancer, colorectal cancer and breast cancer risk in this UK cohort, whether comparing risks for supplement users at one recording point with non-users or comparing risk for users at two recording points with never-users or comparing past/new users with never-users. There was evidence that general supplement use was associated with increased smoking-related cancer risk in the full-sample baseline analysis. However, this was not significant for the users at both recording points, but this secondary analysis was limited by lower numbers and therefore lower power. There was evidence, however, that users at both recording points had different characteristics from women who had not used supplements at one or both points.

Descriptive results from these UKWCS analyses support the inverse supplement hypothesis that supplement users lead a healthier lifestyle than non-users, as found in the UK22, 23, 24 and elsewhere.25, 26, 27, 28, 29, 30 In particular, supplement users in the UKWCS had on average a higher intake of fruit and vegetables and a lower intake of meat and alcohol, and also spent more time exercising vigorously, and had lower body mass indices. Moreover, the results show a trend toward these healthier behaviours from never-users, to new/past users to users at both recording points. Similarly, women taking supplements at all three recording points in the EPIC-Heidelberg cohort in Germany, classed as ‘consistent users’, had the highest intake of dairy products, fish, fruit and vegetables and wine, and the highest physical activity, but the lowest intake of meat compared with the other categories, producing trends for all but physical activity.13 In the cohort of Japanese women, ‘consistent users’ had lower body mass indices and consumed significantly larger amounts of fruit, folate and vitamin C than the other categories, but were more likely to be regular alcohol drinkers and exercised less; in addition, there were no significant trends for green vegetables, meat and fish intake.14 In these previous studies, unlike the current study, ex-smokers were more likely to be inconsistent users. Trends in HRT or contraceptive pill use were found in the current study, but were not examined in the previous studies.

The increased risk in smoking-related cancers in the baseline analysis, found after adjustment for confounders, may relate to evidence that high-dose β-carotene is associated with increased lung4,5 and other smoking-related cancer risks in smokers.31 Conversely, β-carotene has also been inversely associated with smoking-related cancer risks in non-smokers.31 Potentially, β-carotene may act as a pro-oxidant or antioxidant depending upon the biological environment.32 Unfortunately, the amount of β-carotene taken in supplements by the UKWCS users in the current analyses was not reported by the questionnaire and is unknown, although it is usually present in low doses in popular multivitamin supplements.33 Furthermore, as only 11% of the women smoked, the numbers were considered too low to power sub-analyses by smoking status in relation to general supplement use. Alternatively, the increased risks found in the main analysis may be due to other micronutrients in supplements, or due to confounding or multiple testing. The lack of association for smoking-related cancers for users at the two recording points in the UKWCS does not support the main analysis finding; however, this may be due to fewer cases and shorter follow-up time from use at the second survey in this secondary analysis.

There was no evidence of significant associations between general supplement use at baseline and later incidences of total cancers, colorectal cancer or breast cancer for all women, or when explored by menopausal status in the breast cancer analysis. This is in line with results from a breast cancer meta-analysis,34 and also results from two large cohort studies that reported no associations between multivitamin use in women and total cancers, colorectal or breast cancer, and lung cancer incidence.9,10 There is some prior evidence, however, that folic acid in multivitamins is associated with reduced risk of early stages of colorectal cancer.35 The breast cancer results support those from two Danish and US case–control studies on general supplement use;36,37 however, they are in contrast to a Taiwanese case–control study in which general supplement use was associated with a reduced risk for breast cancer (odds ratio=0.40, 95% CI: 0.3, 0.7).38 Selection and recall bias, which can occur in case–control studies, possible lower dietary intakes of micronutrients in the Taiwanese women or the nature of the supplements taken could account for discrepancies in results.

In addition, there were no associations between reporting supplement use at two recording points (both baseline and second survey) and total cancer, colorectal cancer or breast cancer risk for all women, and by menopausal status in the breast cancer analysis. The current study is the first to analyse cancer risks for UK women reporting supplement use at more than one time point in comparison with never-users. Previously, only two studies had analysed general supplement use at more than one recording point in relation to cancer, reporting no associations for ‘consistent users’ in relation to total cancer and major site-specific cancers in a Japanese cohort incidence, and total mortality in the EPIC-Heidelberg cohort in Germany.14,15

Major strengths of the UKWCS study are the prospective design and the large number of cancer cases, particularly in the baseline analysis. However, the power to detect associations was substantially reduced in the analysis of users at two recording points, although numbers in this category were larger (7010 (54%)) compared with previous studies (German men and women: 3559 (18%); Japanese women: 1962 (5.8%)).14,15 Another limitation of this secondary analysis in the UKWCS is that within the category of ‘users at both’ it was not possible to distinguish between regular users who took any supplements for several days each week and sporadic users. Misclassification of sporadic users in the UKWCS into this ‘users at both’ category for premenopausal women, for instance, would have attenuated risks in this category, if cancer risk were lower for sporadic users. However, the questions relating to regular intake, used to define consistent use in the two previous studies,14,15 indicate that the women in this group in these studies were more likely to be stable longer-term users. The average period between baseline and final supplement use recordings was also shorter in the UKWCS. A limitation in the current and previous studies is that it was unknown whether the same types of supplements were taken at baseline as those at second survey: it was unknown for how long, how frequently, how many and at what doses supplements were taken.

Although supplement use in nutrient-deficient populations may be required to reduce the development of cancer,39 any protective effects of general supplement use per se on cancer risk in well-nourished populations such as the UKWCS would be more difficult to explain biologically. As previously reported, many in the UKWCS took a variety of supplement types.12 In line with findings from a small 2008 UK national survey,40 at the UKWCS second survey cod liver oil was more likely to be taken by women of postmenopausal age, whereas multivitamins were more likely to be used by younger women. Whether or not these supplement types have different associations with breast cancer risk, differences in estimates between menopausal statuses in the current analyses were not significant, nor were the increased risk estimates across use categories for premenopausal women significant. Further research is needed into premenopausal users, particularly with larger sample sizes, as research is lacking in this area.

Although observational data such as these are only able to highlight links between diet and disease incidence rather than provide causal evidence, the apparent lack of benefit associated with general supplement use on cancer risks shown in this UK cohort lends support to the guidelines produced by the World Cancer Research Fund that supplement-taking is not advised for reducing cancer risk.

References

  1. 1

    Bates B, Lennox A, Swan G . The National Diet & Nutrition Survey: Headline results from Year 1 of the Rolling Programme (2008/2009). FSA: London, 2009.

  2. 2

    Neuhouser ML, Patterson RE, Levy L . Motivations for using vitamin and mineral supplements. J Am Diet Assoc 1999; 99: 851–854.

  3. 3

    WCRF/AICR. Food, Nutrition, Physical Activity, and the Prevention of Cancer: a Global Perspective. AICR: Washington, DC, 2007.

  4. 4

    Albanes D, Heinonen OP, Taylor PR, Virtamo J, Edwards BK, Rautalahti M et al. Alpha-tocopherol and beta-carotene supplements and lung cancer incidence in the alpha-tocopherol, beta-carotene cancer prevention study: effects of base-line characteristics and study compliance. J Natl Cancer Inst 1996; 88: 1560–1570.

  5. 5

    Omenn GS, Goodman GE, Thornquist MD, Balmes J, Cullen MR, Glass A et al. Effects of a combination of beta carotene and vitamin a on lung cancer and cardiovascular disease. N Engl J Med 1996; 334: 1150–1155.

  6. 6

    Macpherson H, Pipingas A, Pase MP . Multivitamin-multimineral supplementation and mortality: a meta-analysis of randomized controlled trials. Am J Clin Nutr 2013; 97: 437–444.

  7. 7

    Gaziano JM, Sesso HD, Christen WG, Bubes V, Smith JP, MacFadyen J et al. Multivitamins in the prevention of cancer in men: The Physicians' Health Study II randomized controlled trial. JAMA 2012; 308: 1871–1880.

  8. 8

    Hercberg S, Czernichow S, Galan P . Antioxidant vitamins and minerals in prevention of cancers: lessons from the SU.VI.MAX study. Br J Nutr 2006; 96: S28–S30.

  9. 9

    Neuhouser ML, Wassertheil-Smoller S, Thomson C, Aragaki A, Anderson GL, Manson JE et al. Multivitamin use and risk of cancer and cardiovascular disease in the Women's Health Initiative Cohorts. Arch Intern Med 2009; 169: 294–304.

  10. 10

    Park S-Y, Murphy SP, Wilkens LR, Henderson BE, Kolonel LN . Multivitamin use and the risk of mortality and cancer incidence. Am J Epidemiol 2011; 173: 906–914.

  11. 11

    Slatore CG, Littman AJ, Au DH, Satia JA, White E . Long-term use of supplemental multivitamins, vitamin C, vitamin E, and folate does not reduce the risk of lung cancer. Am J Respir Crit Care Med 2008; 177: 524–530.

  12. 12

    Hutchinson J, Burley VJ, Greenwood DC, Thomas JD, Cade JE . High-dose vitamin C supplement use is associated with self-reported histories of breast cancer and other illnesses in the UK Women’s Cohort Study. Public Health Nutr 2011; 14: 768–777.

  13. 13

    Li K, Kaaks R, Linseisen J, Rohrmann S . Consistency of vitamin and/or mineral supplement use and demographic, lifestyle and heath-status predictors: findings from the European Prospective Investigation into Cancer and Nutrition (EPIC)-Heidelberg cohort. Br J Nutr 2010; 104: 1058–1064.

  14. 14

    Hara A, Sasazuki S, Inoue M, Shimazu T, Iwasaki M, Sawada N et al. Use of vitamin supplements and risk of total cancer and cardiovascular disease among the Japanese general population: a population-based survey. BMC Public Health 2011; 11: 540.

  15. 15

    Li K, Kaaks R, Linseisen J, Rohrmann S . Vitamin/mineral supplementation and cancer, cardiovascular, and all-cause mortality in a German prospective cohort (EPIC-Heidelberg). Eur J Nutr 2011; 51: 407–413.

  16. 16

    Batty GD, Kivimaki M, Gray L, Davey Smith G, Marmot MG, Shipley MJ . Cigarette smoking and site-specific cancer mortality: testing uncertain associations using extended follow-up of the original Whitehall study. Ann Oncol 2008; 19: 996–1002.

  17. 17

    Vineis P, Alavanja M, Buffler P, Fontham E, Franceschi S, Gao YT et al. Tobacco and cancer: recent epidemiological evidence. J Natl Cancer Inst 2004; 96: 99–106.

  18. 18

    Cade JE, Burley VJ, Greenwood DC The UKWCS Steering Group. The UK Women's Cohort Study: comparison of vegetarians, fish-eaters and meat-eaters. Public Health Nutr 2004; 7: 871–878.

  19. 19

    Cade JE, Burley VJ, Greenwood DC, The UKWCS Steering Group. Dietary fibre and risk of breast cancer in the UK Women's Cohort Study. Int J Epidemiol 2007; 36: 431–438.

  20. 20

    Taylor EF, Burley VJ, Greenwood DC, Cade JE . Meat consumption and risk of breast cancer in the UK Women's Cohort Study. Br J Cancer 2007; 96: 1139–1146.

  21. 21

    Cade JE, Burley VJ, Greenwood DC . Dietary fibre and risk of breast cancer in the UK Women's Cohort Study. Int J Epidemiol 2007; 36: 431–438.

  22. 22

    Harrison RA, Holt D, Pattison DJ, Elton PJ . Are those in need taking dietary supplements? a survey of 21,923 adults. Br J Nutr 2004; 91: 617–623.

  23. 23

    Kirk SFL, Cade JE, Barrett JH, Conner M . Diet and lifestyle characteristics associated with dietary supplement use in women. Public Health Nutr 1999; 2: 69–73.

  24. 24

    McNaughton SA, Mishra GD, Paul AA, Prynne CJ, Wadsworth MEJ . Supplement use is associated with health status and health-related behaviors in the 1946 British birth cohort. J Nutr 2005; 135: 1782–1789.

  25. 25

    Brownie S . Characteristics of older dietary supplement users: review of the literature. Australas J Ageing 2005; 24: 77–87.

  26. 26

    Frank E, Bendich A, Denniston M . Use of vitamin-mineral supplements by female physicians in the United States. Am J Clin Nutr 2000; 72: 969–975.

  27. 27

    Lyle BJ, Mares-Perlman JA, Klein BEK, Klein R, Greger JL . Supplement users differ from nonusers in demographic, lifestyle, dietary and health characteristics. J Nutr 1998; 128: 2355–2362.

  28. 28

    Patterson RE, Neuhouser ML, White E, Hunt JR, Kristal AR . Cancer-related behavior of vitamin supplement users. Cancer Epidemiol Biomarkers Prev 1998; 7: 79–81.

  29. 29

    Reinert A, Rohrmann S, Becker N, Linseisen J . Lifestyle and diet in people using dietary supplements. a German cohort study. Eur J Nutr 2007; 46: 165–173.

  30. 30

    Shikany JM, Patterson RE, Agurs-Collins T, Anderson G . Antioxidant supplement use in Women's Health Initiative participants. Prev Med 2003; 36: 379–387.

  31. 31

    Touvier M, Kesse E, Clavel-Chapelon F, Boutron-Ruault M-C . Dual association of β-carotene with risk of tobacco-related cancers in a cohort of french women. J Nati Cancer Inst 2005; 97: 1338–1344.

  32. 32

    Palozza P, Serini S, Trombino S, Lauriola L, Ranelletti FO, Calviello G . Dual role of β-carotene in combination with cigarette smoke aqueous extract on the formation of mutagenic lipid peroxidation products in lung membranes: dependence on pO2. Carcinogenesis 2006; 27: 2383–2391.

  33. 33

    Tanvetyanon T, Bepler G . Beta-carotene in multivitamins and the possible risk of lung cancer among smokers versus former smokers. Cancer 2008; 113: 150–157.

  34. 34

    Chan ALF, Leung HWC, Wang SF . Multivitamin supplement use and risk of breast cancer: a meta-analysis. Ann Pharmacother 2011; 45: 476–484.

  35. 35

    Lee JE, Willett WC, Fuchs CS, Smith-Warner SA, Wu K, Ma J et al. Folate intake and risk of colorectal cancer and adenoma: modification by time. Am J Clin Nutr 2011; 93: 817–825.

  36. 36

    Ewertz M, Gill C . Dietary factors and breast cancer risk in Denmark. Int J Cancer 1990; 46: 779–784.

  37. 37

    Moorman PG, Ricciuti MF, Millikan RC, Newman B . Vitamin supplement use and breast cancer in a North Carolina population. Public Health Nutr 2001; 4: 821–827.

  38. 38

    Lee MM, Chang IYH, Horng CF, Chang JS, Cheng SH, Huang A . Breast cancer and dietary factors in Taiwanese women. Cancer Causes Control 2005; 16: 929–937.

  39. 39

    Qiao Y-L, Dawsey SM, Kamangar F, Fan J-H, Abnet CC, Sun X-D et al. Total and cancer mortality after supplementation with vitamins and minerals: follow-up of the Linxian General Population Nutrition Intervention Trial. J Natl Cancer Inst 2009; 101: 507–518.

  40. 40

    GfK Social Research. Consumer consumption of vitamin and mineral food supplements: Random Location Omnibus Survey 2008. London, UK.

Download references

Acknowledgements

The UK Women’s Cohort Study was conceived and designed by JEC and the creation was funded by the World Cancer Research Fund. JH conducted the analysis for this manuscript, wrote the first version and contributed to all other versions. DCG provided statistical advice. All authors contributed to the design of the analysis and to the write-up.

Author information

Correspondence to J Hutchinson.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Hutchinson, J., Burley, V., Greenwood, D. et al. General supplement use, subsequent use and cancer risk in the UK Women’s Cohort Study. Eur J Clin Nutr 68, 1095–1100 (2014). https://doi.org/10.1038/ejcn.2014.85

Download citation

Further reading