European Journal of Clinical Nutrition (2010) 64, 12351238

& 2010 Macmillan Publishers Limited All rights reserved 0954-3007/10

www.nature.com/ejcn

SHORT COMMUNICATION
Serum levels of folate, lycopene, b-carotene, retinol
and vitamin E and prostate cancer risk
J Beilby1, GL Ambrosini2, E Rossi1, NH de Klerk3 and AW Musk2,4

1
Biochemistry Section, PathWest, Queen Elizabeth II Medical Centre, Nedlands, WA, Australia; 2School of Population Health,
University of Western Australia, Nedlands, WA, Australia; 3Telethon Institute for Child Health Research and Centre for
Child Health Research, University of Western Australia, Nedlands, WA, Australia and 4Department of Respiratory Medicine, Sir
Charles Gairdner Hospital, Nedlands, WA, Australia

Previous studies relating increased serum levels of folate and fat-soluble vitamins to prostate cancer risk have variously shown
null associations or to either decrease or increase the risk of developing prostate cancer. Prospective studies of serum folate levels
have been reported to show a null association and increased serum levels to either decrease or increase the risk of subsequently
developing prostate cancer. Similarly, serum b-carotene and lycopene levels have either been reported to be inversely correlated
or not associated with prostate cancer risk. Using a prospective nested case-control study design, which minimized the
possibility of disease effects on serum-vitamin concentrations, we report null associations for serum concentrations of folate,
lycopene, b-carotene, vitamin A and vitamin E, and subsequent development of prostate cancer.
European Journal of Clinical Nutrition (2010) 64, 12351238; doi:10.1038/ejcn.2010.124; published online 4 August 2010

Keywords: folate; lycopene; b-carotene; retinol; vitamin E; prostate cancer

Introduction progression of prostate cancer has also been linked to

Evidence supporting the protective effects of higher fruit
and vegetable consumption on the risks of many cancers
is extensive but previous studies relating serum levels of
folate and fat-soluble vitamins such as b-carotene and vitamin
E to prostate cancer risk have proven equivocal. Prospective
studies of serum folate levels have been reported to variously
show a null association (Weinstein et al., 2003), and increased
serum levels to either decrease (Rossi et al., 2006) or increase
(Hultdin et al., 2005) the risk of subsequently developing
prostate cancer. Similar ambivalent findings have been
reported for b-carotene and lycopene concentrations.
Folate is necessary for DNA synthesis, repair and methyl-
ation and a low folate concentration has been implicated
as a potential promoter of carcinogenesis for example, in
colorectal and lung malignancies (Weinstein et al., 2003).
a possible role in prostate cancer is less certain, hence,
folate concentrations were measured. The initiation and
oxidative stress. Dietary carotenoids are antioxidants
hypothesized to be part of the host defence, potentially
reducing the risk of prostate cancer (Wu et al., 2004).
b-carotene and lycopene occur mainly in orange root
vegetables and tomato-based products respectively, are two
of the most commonly reported carotenoids. Retinol and
vitamin E were also included as diet-derived antioxidants.
The study subjects were drawn from men previously
exposed to blue asbestos who were already participating
in a cancer prevention trial. This was a chemoprevention
programme commenced in 1990, which examined the
efficacy of vitamin supplements in reducing the risk of
malignant mesothelioma, lung cancer and other cancers
(Musk et al., 1998). We conducted a prospective nested
case -control study of serum folate, lycopene, b-carotene,
vitamin A and vitamin E concentrations and prostate cancer
risk among those suffering with subsequent incident prostate
cancer and randomly selected age-matched controls.

Correspondence: Dr E Rossi, Biochemistry Section, PathWest, Queen Elizabeth Subjects

II Medical Centre, Nedlands, Western Australia 6009, Australia.
E-mail: ric.rossi@health.wa.gov.au
Received 27 December 2009; revised 31 March 2010; accepted 8 June 2010; The cohort has been described in detail by Hansen et al.
published online 4 August 2010 (1993) and Berry et al. (2004). At entry to the programme in
 Serum vitamins and prostate cancer risk
J Beilby et al
1236
1990, all participants provided a smoking, medical and Table 1 Baseline characteristics of cases and controls
occupational histories,and height and weight were recorded.
Baseline characteristics Cases Controls P-valuea
Although initially assigned to take either b-carotene (n 96) (n 225)
(synthetic all-trans) or retinol (retinyl palmitate), prelimi-
nary analyses indicated retinol was superior with respect to Mean s.d. Mean s.d.
reducing mesothelioma incidence, and from 1996 the men
Age at diagnosis (years) 69.8 7.2 69.3 6.7 0.56
were re-assigned to take 7.5 mg of retinol/day, except those
Body mass index (kg/m2) 28.4 3.7 27.5 3.9 0.05
with abnormal liver function, who were supplemented with Serum folate (mg/l) 6.06 4.63 7.00 6.51 0.27
0.4 mg of retinol/day. All participants gave their informed Serum lycopene (mmol/l) 0.26 0.19 0.27 0.18 0.88
consent, and this study was approved by the University of Serum b-carotene (mmol/l) 0.51 0.37 0.60 0.53 0.17
Serum vitamin A (mmol/l) 3.22 0.67 3.20 0.72 0.98
Western Australias Human Research Ethics Committee and
Serum vitamin E (mmol/l) 41.2 12.6 39.1 10.6 0.14
the Clinical Drug Trials Committee of the Sir Charles
Gairdner Hospital. n (%) n (%) P-valueb
Serum vitamins were measured on a non-fasting sample
Smoking category
obtained at entry to the study, before consumption of any Current smoker 11 (11.5) 25 (11.1) 0.93
supplements. Serum folate was analyzed by automated Former smoker 60 (62.5) 146 (64.6)
immunoassay (Architect i2000, Abbott, Chicago, IL, USA) Never smoked 25 (26) 55 (24.3)
and serum lycopene, b-carotene, vitamin A (retinol) and
Abbreviation: s.d., standard deviation.
vitamin E (a-tocopherol) by high performance liquid a
t-test.
chromatography (Thurnham et al., 1988). b 2
w -test.
All programme participants were followed up for incident
prostate cancer diagnosed after at least 1 year of enrolment in No differences in baseline characteristics were seen between
the programme until 31st December 2004, or date of death, cases and controls (Table 1), except mean body mass index
which ever occurred first. Prostate cancer cases were histologi- was slightly higher for cases (P 0.05). Potential confounders
cally confirmed cases identified through record linkage with were selected on the basis of the original trial parameters,
the Western Australian Cancer Registry and the Australian which were related to the risk of malignant mesothelioma
National Cancer Statistics Clearing House. Three controls were and lung cancer. The following potential confounders
randomly selected for each case for the nested case-control (measured at baseline) were investigated: age, body mass
analysis, after matching on age (within 5 years) of the case. index, source of asbestos exposure and smoking status
For some cases, particularly those aged over 70 years, only two (previous smoker, current smoker or never smoked). To
rather than three suitable controls were available for each case. control for randomly assigned vitamin A supplements, an
intention-to-treat analysis was conducted, whereby each
subject was analyzed according to the original supplement
Statistical analyses assigned to them at entry to the program. None of the
Baseline characteristics of cases and controls were compared potential confounders examined were statistically signifi-
using t-tests for continuous variables (age, body mass index cant, nor did they affect the odds ratio of the main effect.
and serum-vitamin concentrations) and w2-tests for catego- However, age and vitamin A supplement were retained in
rical variables (smoking status). Prostate cancer risk was all models.
analyzed using conditional logistic regression models in Table 2 shows prostate cancer risk analyzed using condi-
Stata software (Intercooled Stata 9.0 for Windows. 9.0 ed. tional logistic regression with serum-vitamin concentrations
College Station, Texas, USA: StataCorp LP; 2005). Serum as tertiles. Odds ratios and 95% confidence intervals were
vitamin concentrations were analyzed as tertiles (33rd and calculated for each tertile relative to the tertile with the
67th percentiles), using the xtile command in Stata and on lowest vitamin concentration. The second tertile of lycopene
the basis of the distributions in controls. Vitamin concentra- levels suggested a reduction in risk, but this did not continue
tions were also analyzed as continuous variables after log into the third tertile. Levels displayed a trend towards an
transformation, because of non-normal distributions. The association with protection but did not achieve significance.
level for significance was set to 5% for all statistical tests.
The study had 80% power to detect as significant at the 0.05
levels, a 25% increase in serum lycopene or serum folate Discussion
(after log transformation) in controls compared with cases.
Serum b-carotene levels have either been reported to be
inversely correlated with prostate cancer risk in a case-
Results control study (Chang et al., 2005) or not associated in a large
prospective nested case-control study (Gann et al., 1999).
A total of 96 cases of prostate cancer were identified during Similar controversy applies to serum lycopene levels, which
the follow-up period, and matched with 226 controls. were reported by Chang et al. (2005) to have no association

European Journal of Clinical Nutrition

 Serum vitamins and prostate cancer risk
J Beilby et al
1237
Table 2 Serum vitamin concentrations and prostate cancer riska

Serum vitamin Tertiles of plasma vitamin P for trend Continuous ORb

1 2 3

Folate
Cases (non-cases) 33 (70) 32 (70) 27 (74)
Minmax (mg/l) 1.503.80 3.906.20 6.3045.1
OR (95% CI) 1 1.10 (0.572.11) 1.09 (0.482.46) 0.83 0.99 (0.561.77)

Lycopene
Cases (non-cases) 38 (65) 31 (91) 25 (53)
Minmax (mmol/l) 00.19 0.200.30 0.311.30
OR (95% CI) 1 0.55 (0.300.99) 0.77 (0.401.47) 0.36 0.98 (0.681.43)

b-carotene
Cases (non-cases) 34 (70) 31 (75) 29 (70)
Minmax (mmol/l) 0.100.30 0.310.60 0.613.70
OR (95% CI) 1 0.86 (0.481.54) 0.83 (0.451.55) 0.55 0.79 (0.561.11)

Vitamin A
Cases (non-cases) 29 (67) 25 (69) 41 (84)
Minmax (mmol/l) 1.102.88 2.903.30 3.355.79
OR (95% CI) 1 0.80 (0.421.52) 1.12 (0.602.08) 0.66 1.09 (0.343.54)

Vitamin E
Cases (non-cases) 25 (72) 28 (73) 42 (75)
Minmax (mmol/l) 21.033.4 33.640.4 40.893.2
OR (95% CI) 1 1.06 (0.532.10) 1.46 (0.782.73) 0.19 1.76 (0.694.50)

Abbreviations: CI, confidence interval; Max, maximum; Min, minimum; OR, odds ratio.
a
Adjusted for age, administered vitamin A supplement.
b
Serum concentrations analyzed as log transformed, continuous variables.

with prostate cancer risk or by Gann et al. (1999) to be
inversely correlated with risk. A more recent prospective
nested case-control study found a significant inverse associa-
tion between serum lycopene levels only for men over 65
years without a family history of prostate cancer (Wu et al.,
2004).
There is no known connection between asbestos exposure
experienced in both the cases and controls and prostate
cancer risk, nor can we speculate on any changes to
effectiveness of the measured micronutrients in the presence
of asbestos exposure. This case-control study examined
several key fat-soluble vitamins, as well as plasma folate
using a prospective design. The risk of subsequent develop-
ment of prostate cancer was not associated with serum
concentrations of folate, vitamin A, b-carotene or vitamin E.
The prospective study design is an important strength,
which minimized the possibility of disease or treatment
effects on serum vitamin levels. Analysis of plasma vitamin
levels also avoids possible difficulties with poor dietary recall
when subjects complete food frequency questionnaires.
Conclusion
Our study findings are tempered by a relatively small sample
size and recognition that a single sampling of serum
micronutrients does not assess long term status. Risk factors
specific to prostate cancer, such as prostate specific antigen
levels at baseline, race and family history were not available,
however we can state the subjects were predominantly of
European descent.
In conclusion, we report null associations for prostate
cancer and plasma folate, b carotene, vitamin concentrations
similar to results reported by Weinstein et al. (2003)
for folate, by Gann et al. (1999) for b-carotene and by
Chang et al. (2005) for lycopene.
Conflict of interest
The authors declare no conflict of interest.
References
Berry G, de Klerk NH, Reid A, Ambrosini GL, Fritschi L, Olsen NJ et al.
(2004). Malignant pleural and peritoneal mesotheliomas in former
miners and millers of crocidolite at Wittenoom, Western Australia.
Occup and Environ Med 61, e14.
Chang S, Erdman Jr JW, Clinton SK, Vadiveloo M, Strom SS,
Yamamura Y et al. (2005). Relationship between plasma
carotenoids and prostate cancer. Nutr Cancer 53, 127134.
Gann PH, Ma J, Giovannucci E, Willett W, Sacks FM, Hennekens CH
et al. (1999). Lower prostate cancer risk in men with elevated
plasma lycopene levels: results of a prospective analysis. Cancer Res
59, 12251230.

European Journal of Clinical Nutrition

 Serum vitamins and prostate cancer risk
J Beilby et al
1238
Hansen J, de Klerk NH, Eccles JL, Musk AW, Hobbs MS (1993).
Malignant mesothelioma after environmental exposure to blue
asbestos. Int J Cancer 54, 578581.
Hultdin J, Van Guelpen B, Bergh A, Hallmans G, Stattin P (2005).
Plasma folate, vitamin B12, and homocysteine and prostate cancer
risk: a prospective study. Int J Cancer 113, 819824.
Musk AW, de Klerk NH, Ambrosini GL, Eccles JL, Hansen J, Olsen NJ
et al. (1998). Vitamin A and cancer prevention I: observations in
workers previously exposed to asbestos at Wittenoom, Western
Australia. Int J Cancer 75, 355361.
Rossi E, Hung J, Beilby JP, Knuiman MW, Divitini ML, Bartholomew H
(2006). Folate levels and cancer morbidity and mortality: prospec-
tive cohort study from Busselton, Western Australia. Ann Epidemiol
16, 206212.
European Journal of Clinical Nutrition
Thurnham DI, Smith E, Flora PS (1988). Concurrent liquid-
chromatographic assay of retinol, alpha-tocopherol, beta-
carotene, alpha-carotene, lycopene, and beta-cryptoxanthin in
plasma, with tocopherol acetate as internal standard. Clin Chem
34, 377381.
Weinstein SJ, Hartman TJ, Stolzenberg-Solomon R, Pietinen P, Barrett
MJ, Taylor PR et al. (2003). Null association between prostate
cancer and serum folate, vitamin B(6), vitamin B(12), and
homocysteine. Cancer Epidemiol Biomarkers Prev 12 (11 Pt. 1),
12711272.
Wu K, Erdman Jr JW, Schwartz SJ, Platz EA, Leitzmann M, Clinton SK
et al. (2004). Plasma and dietary carotenoids, and the risk of
prostate cancer: a nested case-control study. Cancer Epidemiol
Biomarkers Prev 13, 260269.
Copyright of European Journal of Clinical Nutrition is the property of Nature Publishing Group and its content
may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express
written permission. However, users may print, download, or email articles for individual use.