Content uploaded by Emma Leacy
Author content
All content in this area was uploaded by Emma Leacy on May 19, 2016
Content may be subject to copyright.
University of Dublin, Trinity College, 2014
A Meta-Analysis of Dietary
Carotenoids and Prostate
Cancer Incidence
Emma Leacy – 10314771
Supervised by Dr. Katarina Bälter, Dr. Jennifer Protudjer, and Dr. Arvid Sjölander
at the Department of Medical Epidemiology and Biostatistics, Karolinska
Institutet, Stockholm
A Thesis Submitted in Partial Fulfilment of Module PG4902 for the degree of B.Sc. in
Human Health & Disease
Page | 2
Contents
Declaration and Statement of Plagiarism ....................................................................................................... 6
Acknowledgements ........................................................................................................................................ 7
List of Figures .................................................................................................................................................. 8
List of Tables ................................................................................................................................................... 8
1. Abstract ....................................................................................................................................................... 9
2. Background ...............................................................................................................................................10
2.1. Prostate Cancer Epidemiology ...........................................................................................................10
2.2. Diet and Prostate Cancer ...................................................................................................................11
2.3. Carotenoids ........................................................................................................................................12
2.4. α-Carotene .........................................................................................................................................14
2.5. β-Cryptoxanthin .................................................................................................................................15
2.6. Lutein .................................................................................................................................................15
2.7. Zeaxanthin .........................................................................................................................................16
2.8. Blood Carotenoid Levels ....................................................................................................................18
2.9. Objective ............................................................................................................................................18
3. Materials and Methods .............................................................................................................................19
3.1. Search Strategy/Identification of Literature ......................................................................................19
3.2. Inclusion & Exclusion Criteria.............................................................................................................19
3.3. Data Extraction & Evaluation .............................................................................................................20
3.4. Statistical Analysis ..............................................................................................................................21
4. Results .......................................................................................................................................................22
4.1. Study Characteristics..........................................................................................................................22
4.2. Dietary Assessment ............................................................................................................................22
4.3. Dietary Carotenoid Intake and Prostate Cancer Risk .........................................................................25
4.4. α-Carotene .........................................................................................................................................25
4.5. β-Cryptoxanthin .................................................................................................................................26
4.6. Lutein .................................................................................................................................................27
4.7. Lutein & Zeaxanthin ...........................................................................................................................27
5. Discussion .................................................................................................................................................29
5.1. Measures of Association ....................................................................................................................29
5.2. Study Design ......................................................................................................................................29
5.3. Dietary Assessment ............................................................................................................................30
5.4. Heterogeneity ....................................................................................................................................31
5.5. Exclusion of Jian et al58 ......................................................................................................................32
Page | 3
5.6. Prostate Cancer Definition .................................................................................................................33
5.7. Publication Bias ..................................................................................................................................34
5.8. Limitations of Current Analysis ..........................................................................................................34
5.9. Supplementation ...............................................................................................................................34
5.10. Synergy.............................................................................................................................................35
5.11. Carotenoid Bioavailability ................................................................................................................36
5.12. Interpersonal Differences ................................................................................................................36
5.13. Plant-Based Diets .............................................................................................................................37
5.14. Future Directions: Developing a Prostate Cancer Diet Score (PCDS)...............................................37
6. Conclusion .................................................................................................................................................38
Appendix 1 – Results from Most Recent Meta-analyses of Dietary Factors and Prostate Cancer Incidence
......................................................................................................................................................................39
Appendix 2 – Details of Initial Literature Review of Dietary Factors and Prostate Cancer ..........................40
Appendix 3 – Summary of Dietary Assessment Methods Used in Included Studies ....................................41
Appendix 4 – Exposure Assessment Boundaries and Adjusted Covariates ..................................................45
4.1. α-Carotene .........................................................................................................................................45
4.2. β-Cryptoxanthin .................................................................................................................................47
4.3. Lutein .................................................................................................................................................49
4.4. Lutein and Zeaxanthin .......................................................................................................................50
Appendix 5 – Contents of Do Files Used to Complete Meta-analysis in Stata .............................................51
Figure B - Do File Used for Case-control/NCC study analysis ...................................................................51
Figure C - Do File Used for Cohort/Case-cohort Study Analysis ...............................................................51
Appendix 6 – Full Results of Meta-Analyses of Dietary Intake of Carotenoids and Prostate Cancer
Incidence .......................................................................................................................................................52
6.1. α-Carotene .........................................................................................................................................52
Table A - Full Results of Included Case-Control/NCC Studies ...............................................................52
Figure D - Forrest Plot Showing Risk Estimates from Included Case–Control/NCC Studies .................52
Figure E - Funnel Plot Examining Publication Bias in Included Case-Control/NCC Studies of Dietary α-
Carotene and Prostate Cancer Risk ......................................................................................................53
Table B - Full Results of Included Cohort/Case-Cohort Studies ............................................................53
Figure F - Forrest Plot Showing Risk Estimates from Included Cohort/Case-Cohort Studies of Dietary
α-Carotene Intake and Prostate Cancer Risk ........................................................................................54
Figure G - Funnel Plot Examining Publication Bias in Included Cohort/Case-Cohort Studies of Dietary
α-Carotene and Prostate Cancer Risk ...................................................................................................54
6.2. β-Cryptoxanthin .................................................................................................................................55
Table C - Full Results of Included Case-Control/NCC Studies ...............................................................55
Figure H - Forrest Plot Showing Risk Estimates from Included Case–Control/NCC Studies of Dietary β-
Cryptoxanthin Intake and Prostate Cancer Risk ...................................................................................55
Page | 4
Figure I - Funnel Plot Examining Publication Bias in Included Case-Control/NCC Studies of Dietary β-
Cryptoxanthin and Prostate Cancer Risk ..............................................................................................56
Table D - Full Results of Included Cohort/Case-Cohort Studies ............................................................56
Figure J - Forrest Plot Showing Risk Estimates from Included Cohort/Case-Cohort Studies of Dietary
β-Cryptoxanthin Intake and Prostate Cancer Risk ................................................................................57
Figure K - Funnel Plot Examining Publication Bias in Included Cohort/Case-Cohort Studies of Dietary
β-Cryptoxanthin and Prostate Cancer Risk ...........................................................................................57
6.3. Lutein .................................................................................................................................................58
Table E - Full Results of Included Case-Control/NCC Studies ................................................................58
Figure L - Forrest Plot Showing Risk Estimates from Included Case–Control/NCC Studies of Dietary
Lutein Intake and Prostate Cancer Risk ................................................................................................58
Figure M - Funnel Plot Examining Publication Bias in Included Case-Control/NCC Studies of Lutein
and Prostate Cancer Risk ......................................................................................................................59
Table F - Full Results of Included Cohort/Case-Cohort Studies ............................................................59
Figure M - Forrest Plot Showing Risk Estimates from Included Cohort/Case-Cohort Studies of Dietary
Lutein Intake and Prostate Cancer Risk ................................................................................................60
Figure N - Funnel Plot Examining Publication Bias in Included Cohort/Case-Cohort Studies of Dietary
Lutein and Prostate Cancer Risk ...........................................................................................................60
6.4. Lutein & Zeaxanthin ...........................................................................................................................61
Table G - Full Results of Included Case-Control/NCC Studies ...............................................................61
Figure O - Forrest Plot Showing Risk Estimates from Included Case–Control/NCC Studies of Dietary
Lutein & Zeaxanthin Intake and Prostate Cancer Risk ..........................................................................61
Figure P - Funnel Plot Examining Publication Bias in Included Case-Control/NCC Studies of Dietary
Lutein & Zeaxanthin and Prostate Cancer Risk .....................................................................................62
Table H - Full Results of Included Cohort/Case-Cohort Studies ............................................................62
Figure Q - Forrest Plot Showing Risk Estimates from Included Cohort/Case-Cohort Studies of Dietary
Lutein & Zeaxanthin Intake and Prostate Cancer Risk ..........................................................................63
Figure R - Funnel Plot Examining Publication Bias in Included Cohort/Case-Cohort Studies of Dietary
Lutein & Zeaxanthin and Prostate Cancer Risk .....................................................................................63
Appendix 7 – Meta-Analysis of Case-Control/Nested Case-Control Studies of Dietary Carotenoids and
Prostate Cancer Risk Excluding Results from Jian et al .................................................................................64
7.1. α-Carotene .........................................................................................................................................64
Table I - Full Results of Included Studies ...............................................................................................64
Figure S - Forrest Plot Showing Risk Estimates from Included Studies of Dietary α-Carotene Intake
and Prostate Cancer Risk ......................................................................................................................64
Figure T - Funnel Plot Examining Publication Bias in Included Studies of Dietary α-Carotene and
Prostate Cancer Risk .............................................................................................................................65
7.2. β-Cryptoxanthin .................................................................................................................................65
Table J - Full Results of Included Studies ..............................................................................................65
Page | 5
Figure U - Forrest Plot Showing Risk Estimates from Included Studies of Dietary β-Cryptoxanthin
Intake and Prostate Cancer Risk ...........................................................................................................66
Figure V - Funnel Plot Examining Publication Bias in Included Studies of Dietary β-Cryptoxanthin and
Prostate Cancer Risk .............................................................................................................................66
7.3. Lutein & Zeaxanthin ...........................................................................................................................67
Table K - Full Results of Included Studies..............................................................................................67
Figure W - Forrest Plot Showing Risk Estimates from Included Studies of Dietary lutein & Zeaxanthin
Intake and Prostate Cancer Risk ...........................................................................................................67
Figure X - Funnel Plot Examining Publication Bias in Included Studies of Dietary lutein & Zeaxanthin
and Prostate Cancer Risk ......................................................................................................................68
References ....................................................................................................................................................69
Page | 6
Declaration and Statement of Plagiarism
I declare that this thesis has not been submitted as an exercise for a degree at this or any other university
and is entirely my own work.
I agree to deposit this thesis in the University’s open access institutional repository or allow the library to
do so on my behalf, subject to Irish Copyright Legislation and Trinity College Library conditions of use and
acknowledgement.
Signed Date
___________________________________________________ ________________________
Page | 7
Acknowledgements
There are a number of people I would like to thank at MEB who facilitated me in carrying out this
project. Firstly, I would like to thank Dr. Katarina Bälter for allowing me to complete my degree project at
Karolinska Institutet. You, and everybody else in the research group made me feel very welcome and
made my time there very enjoyable. To my co-supervisors Jennifer Protudjer and Arvid Sjölander – thank
you for all your help and invaluable knowledge and guidance throughout this project. And also Camilla
Wiklund for always brightening up the office.
My family have been incredibly supportive throughout my time in Stockholm, particularly Paul
Leacy and Sinead Mac. Thanks to mam, dad and the kids for always being there and encouraging me to
do my best. My friends have also been incredibly helpful and enthusiastic while I’ve been away, and they
never fail to put a smile on my face.
To those mentioned and everybody else who has helped me through these last three months –
tack så mycket!
The doctor of the future will give no medicine, but will interest her or his
patients in the care of the human frame, in a proper diet, and in the cause
and prevention of disease.
Thomas Edison, US inventor (1847 - 1931)
Page | 8
List of Figures
Figure 1 – Relation between Prevalence of Prostate Cancer at Autopsy, Clinically Diagnosed Prostate
Cancer, and Prostate Cancer Deaths
Figure 2 – Disparities in Prostate Cancer Incidence Worldwide
Figure 3 – Classification of Carotenoids
Figure 4 – Differences in Structure of Lutein and Zeaxanthin
Figure 5 – PRISMA Flow Diagram Demonstrating the Process of Study Selection
Figure 6 – Forrest Plot for Meta-analysis of Case-control/Nested Case-control Studies of α-Carotene and
Risk of Prostate Cancer
Figure 7 – Funnel Plot for Test of Publication Bias in Case-control/Nested Case-control Studies of α-
Carotene and Prostate Cancer
Figure 8 – Forrest Plot for Meta-analysis of Case-control/Nested Case-control Studies of β-Cryptoxanthin
and Prostate Cancer
Figure 9 – Forrest Plot for Meta-analysis of Case-control/Nested Case-control Studies of Lutein and Risk of
Prostate Cancer
Figure 10 – Forrest Plot for Meta-analysis of Case-control/Nested Case-control Studies of Lutein &
Zeaxanthin and Risk of Prostate Cancer
List of Tables
Table 1 – Top Food Sources of α-Carotene, taken from USDA Food Composition Database
Table 2 – Top Food Sources of β-Cryptoxanthin, taken from USDA Food Composition Database
Table 3 – Top Food Sources of Lutein & Zeaxanthin, taken from USDA Food Composition Database
Table 4 – Characteristics of Case-Control Studies of Dietary Intake of Carotenoids Included in Meta-
analysis
Table 5 – Characteristics of Cohort/Case-Cohort Studies of Dietary Intake of Carotenoids Included in
Meta-analysis
Table 6 – Summary of Results of Meta-analysis of Four Carotenoids and Risk of Prostate Cancer
Page | 9
1. Abstract
Background: Prostate cancer is one of the most common cancers in the world. However, disparities in
incidence rates worldwide have suggested that lifestyle factors, particularly diet, may play a role in its
development. Carotenoids have exhibited multiple anti-cancer effects, and increased intakes of high-
carotenoid foods have been shown to be protective against prostate cancer in epidemiological
investigations. The aim of this project was to complete a meta-analysis of dietary intake of four
carotenoids – α-carotene, β-cryptoxanthin, lutein, and zeaxanthin – to determine their role in prostate
cancer incidence.
Methods: A PubMed literature search and a systematic review of the literature was performed to identify
studies measuring carotenoid intake and prostate cancer risk. Estimates of OR or HR for highest versus
lowest categories of intake were pooled for each individual carotenoid for case-control/NCC studies and
cohort/case-cohort studies, respectively. Tests for heterogeneity and publication bias were also carried
out.
Results: A total of sixteen published articles were included in the analysis. A significantly reduced risk of
prostate cancer was found for higher intakes of each of the four carotenoids in case-control/NCC studies,
but not for cohort/case-cohort studies. Pooled ORs for lutein (0.76, 95% CI = 0.60-0.97, p = 0.03) and
lutein & zeaxanthin (0.82, 95% CI = 0.75-0.89, p = 0.00) showed the strongest risk reductions, while α-
carotene (OR = 0.92, 95% CI = 0.84-1.00, p = 0.04) and β-cryptoxanthin (OR = 0.91, 95% CI = 0.83-0.99, p =
0.03) showed more modest protective effects. Cohort/case-cohort studies also expressed reduced risks
for higher intakes (lutein showed no association; HR = 1.00, 95% CI = 0.91-1.10, p = 0.97), though these
results were not statistically significant. No publication bias was detected, though there was significant
heterogeneity between included studies.
Conclusion: There appears to be an inverse association for intake of α-carotene, β-cryptoxanthin, lutein,
and zeaxanthin and prostate cancer. Increased intakes of high carotenoid foods may be protective
against prostate cancer development.
Page | 10
Figure 1 - Relation Between Prevalence of Prostate
Cancer at Autopsy, Clinically Diagnosed Prostate
Cancer, and Prostate Cancer Deaths. Adapted from
Damber and Aus9
2. Background
2.1. Prostate Cancer Epidemiology
Prostate cancer accounts for 15% of all cancer diagnosed in men, with over 1.1 million cases in
20121. Prostate cancer is the most common cancer in Europe, with an estimated incidence rate of 96.0
per 100,0002. Sweden and Ireland have the 3rd and 4th highest rates in Europe, with 175.2 and 168.7 per
100,000 respectively. These high incidence rates have been attributed to an increased prevalence of
prostate-specific antigen (PSA) testing, leading to higher diagnoses of non-fatal prostate cancers3.
Mortality rates have remained relatively low (307,000 deaths worldwide in 2012), despite a high
prevalence of latent prostate cancer at all ages (Figure 1). Autopsy studies have suggested that the
majority of men over 85 years of age have some degree of histological prostate cancer4.
Cancer of the prostate generally occurs in
men over the age of 65, and there is an increased risk
for men whose family have history of the disease5.
Initiation of prostate cancer involves downregulation
of multiple molecular signalling pathways, though
androgen receptor signalling is key for progression to
aggressive forms6. A number of genetic factors have
been identified, and epigenetic mechanisms are
emerging as a candidate for novel treatments7.
Benign prostatic hyperplasia (BPH) is another
common condition of the prostate which shares
many pathophysiological traits with prostate cancer8.
Despite being the second most common cancer among men, there exists substantial worldwide
variance in prostate cancer incidence, with developed countries accounting for almost 70% of cases1. The
highest incidence rates are reported in Australia & New Zealand, North America, and Northern and
Western Europe, with South-East & South-Central Asian regions reporting the lowest rates. Rates in Asia
are almost six times lower than their ‘Western’ counterparts10. Disparities also exist between prostate
cancer incidence and mortality worldwide (Figure 2). Furthermore migrants from low- to high-risk regions
experience increased incidence and mortality from prostate cancer within two generations11. Taken
together, these trends are highly suggestive of the influence of environmental factors upon prostate
cancer risk.
As the old adage goes; prevention is better than cure. The key aim of cancer chemoprevention
studies of dietary constituents is to identify active ingredients and explicate their underlying mechanisms.
This can aid in designing strategies for intervention trials, and ultimately educate people on the best ways
Page | 11
Figure 2 - Disparities in Prostate Cancer Incidence
Worldwide1
to avoid illness and improve their overall health. Prostate cancer is an ideal candidate for
chemoprevention studies. The high incidence rates and long latency period mean there is a large
therapeutic window for dietary intervention treatment. Identifying conclusive dietary factors which are
protective or damaging to prostate cancer development will create a simple, effective, and low-cost
method for reducing disease prevalence worldwide.
A recent British study determined that almost
43% of cancers are attributed to lifestyle factors. In the
context of prostate cancer, smoking13 and obesity14
are associated with higher risks of fatal prostate
cancer (30% and 15%, respectively). A recent study
carried out by this research group at Karolinska15
showed that higher levels of physical activity were
associated with decreased rates of prostate cancer
mortality, with a hazard ratio of 0.62 (95% CI = 0.41-
0.94) for men who walked/bicycled 20-60 minutes per
day compared to men in the lowest category.
2.2. Diet and Prostate Cancer
One lifestyle factor that has received considerable focus in relation to prostate cancer is diet.
There have been a number of epidemiological investigations into the influences of different foods and
prostate cancer risk. Although an analysis of the EPIC cohort16 (European Prospective Investigation into
Cancer and Nutrition) found no association between fruit and vegetable intake and prostate cancer
incidence, analysis of specific vegetables and groups have found significant associations.
Several analyses of different fruits and vegetable classes have been completed, but there is
considerable variance in how studies classify vegetables into specific groups. Notable studies involve the
analysis of effects of legume and cruciferous vegetable intake. Legumes have been widely examined due
to the influence of soy foods and soy isoflavones in prostate cancer, which have demonstrated significant
protective effects17. Cruciferous vegetables, particularly those from the Brassica family, have a significant
protective effect against prostate cancer in both epidemiologic and laboratory studies18. In a recent
meta-analysis by Liu et al19, high consumers of cruciferous vegetables had a relative risk of 0.90 (95% CI =
0.85-0.96) for prostate cancer compared to low consumers. Though analysis of the effects of vegetables
can be beneficial, these analyses do not reflect the independent influences of foods (or nutrients) within
their groupings.
Page | 12
There have been a number of meta-analyses of different dietary factors and their effect on
prostate cancer incidence (Appendix 1). Fish20, cruciferous vegetables19, coffee21, and total soy food17
have all been shown to significantly decrease prostate cancer risk, while dairy products22 and total fat23
increase risk. Similarly, food constituents have been examined to determine their influence on the
disease. Daidzein and genistein, two common isoflavones found in soy, were found to decrease prostate
cancer incidence when consumed at higher amounts17.
Of the completed meta-analyses that reached statistical significance, many of them have
examined foods which are high in carotenoids. Carotenoids have exhibited multiple protective effects
against cancer (see sections 2.3-2.7), and are commonly consumed by many cultures. The most recent
meta-analysis on this topic related to carrot intake and prostate cancer risk24. The analysis of 10 studies
showed a significantly decreased risk of prostate cancer (OR = 0.82, 95% CI = 0.70-0.97) for men with high
compared to low carrot intakes. This study also found a dose-response association between carrot
consumption and reduced risk of prostate cancer. An increase of one serving of carrots per week yielded
a risk estimate of 0.95 (95% CI = 0.90-0.99), and for each 10g per day increase this estimate was 0.96
(95% CI = 0.94-0.99). Carrots are particularly high in α-carotene, and many other foods which have been
shown to be protective against prostate cancer contain high concentrations of carotenoids.
2.3. Carotenoids
Carotenoids are fat-soluble organic pigments that are found in the chloroplasts and
chromoplasts of plants. The name “carotene” actually comes from the Latin carota, meaning ‘carrot’.
Carotenoids are responsible for the bright, orange-hued colours of many foods, and some of them are
used as food colourings.
Figure 3 - Classification of Carotenoids. Those shown in yellow have vitamin A activity.
Carotenoids
Carotenes
α-Carotene
Xanthophylls
β-
Cryptoxanthin Lutein Zeaxanthin
Page | 13
Over 700 carotenoids have been identified to date, though only a fraction of these are present in
human diets25. They are split into two classes (Figure 3); xanthophylls (which contain oxygen) and
carotenes (which do not contain oxygen). Carotenoids are biosynthesised from common precursors
(geranylgeranyl pyrophosphate) and a number of enzymes facilitate their production. Humans cannot
synthesise carotenoids themselves and must acquire them entirely from dietary sources. People with
diets rich in carotenoids are at a lower risk for cardiovascular and ocular diseases, as well as many
cancers26. The health benefits of carotenoids may be in part explained by their relationship with vitamin
A. Vitamin A is essential for a variety of biological processes, including organogenesis, immune
competence, tissue differentiation, and the visual cycle27. Some carotenoids have provitamin A activity,
and their relationships with vitamin A may synergistically contribute to their overall health benefits.
The most prevalent dietary carotenoids are α-carotene, β-carotene, lycopene, lutein, zeaxanthin,
and β-cryptoxanthin28. However, because they are not classified as essential nutrients, values for
recommended daily intake have not been established29. Carotenoids are found in a wide variety of foods,
most commonly fruits and vegetables. Top dietary sources for α-carotene, β-cryptoxanthin, and lutein &
zeaxanthin are shown in Tables 1-3 respectively. Hardin et al30 investigated the influence of high intakes
of fruits and vegetables rich in carotenoids (as well as other dietary components) on prostate cancer.
High carotenoid fruits encompassed apricots, orange juice, grapefruit juice, peaches, nectarines, plums,
cantaloupe, orange melon, mango, oranges, grapefruits, and tangerines. High carotenoid vegetables
included in the analysis were beans, fresh tomatoes, tomato juice, vegetable juice, broccoli, cauliflower,
cabbage, Brussels sprouts, green peas, vegetable, minestrone, and tomato soups, carrots, green salad,
winter and summer squash, red peppers/chilies, yams/sweet potatoes, spinach, mustard greens, and
collards.
Lycopene is the most abundant carotenoid found in blood and has already been extensively
studied as a possible chemopreventative tool for prostate cancer. Having demonstrated multiple anti-
cancer mechanisms in in-vitro studies31, investigations into lycopene intake have shown a protective
effect against prostate cancer32 (OR for higher intakes compared to lower = 0.93; 95% CI = 0.86-1.01).
High intakes of tomatoes (the most abundant source of lycopene) have also led to reduced risks
compared to lower intakes (OR = 0.81; 95% CI = 0.59-1.10). Another carotenoid that has been highly
investigated in the context of cancer is β-carotene. A large RCT found that β-carotene supplementation
led to an increased rate of cancer at several sites, including a 23% increase in incidence and 15% increase
in mortality from prostate cancer33. The World Cancer Research Fund Report investigated the effect of
serum, dietary, and supplemental β-carotene on prostate cancer, and determined that neither β-
carotene nor foods containing it are likely to have a substantial effect on the risk of prostate cancer34.
Page | 14
Table 1 – Top Food Sources of α-Carotene, taken from USDA Food Composition Database35
Description
µg/100 g
Carrot, dehydrated
14,251
Peppers, sweet, red, freeze-dried
6,931
Pumpkin, canned, without salt
4,795
Pumpkin, canned, with salt
4,795
Carrot juice, canned
4,342
Pumpkin, raw
4,016
Carrots, cooked, boiled, drained, without salt
3,776
Carrots, cooked, boiled, drained, with salt
3,776
Carrots, baby, raw
3,767
Babyfood, carrots and beef, strained
3,716
Carrots, frozen, cooked, boiled, drained, without salt
3,716
Carrots, frozen, cooked, boiled, drained, with salt
3,716
Carrots, raw
3,477
Babyfood, carrots, toddler
3,340
Carrots, frozen, unprepared
2,958
Soup, cream of vegetable, dry, powder
2,820
Carrots, canned, regular pack, drained solids
2,743
Carrots, canned, no salt added, solids and liquids
2,743
Carrots, canned, no salt added, drained solids
2,743
Pumpkin, cooked, boiled, drained, without salt
2,715
Pumpkin, cooked, boiled, drained, with salt
2,715
Carrots, canned, regular pack, solids and liquids
2,692
Babyfood, vegetables, carrots, junior
2,682
Vegetables, mixed, canned, drained solids
2,636
2.4. α-Carotene
α-carotene is the second most common form of carotene after β-carotene. The two differ only in
structure by the position of a double bond (and consequentially, a hydrogen atom) in the cyclic group at
one end. As it contains a retinyl group (a β-ionone ring which allows an isoprenoid ring to attach), α-
carotene has a small degree of provitamin A activity. Serum α-carotene concentrations were inversely
associated with all-cause mortality, as well as cardiovascular disease, cancer, and all other causes36. α-
carotene was also found to inhibit proliferation of endometrial, mammary, and lung human cancer cells
in culture37.
The major dietary sources of α-carotene are from yellow/orange vegetables (carrots, sweet
potatoes, winter squash) and dark green vegetables (broccoli, green beans, green peas, spinach, turnip
greens, collards, leaf lettuce, avocado). As shown in Table 1, carrots are the top source of α-carotene, and
it is also found in high amounts in other yellow-orange vegetables such as peppers and pumpkin.
Page | 15
2.5. β-Cryptoxanthin
β-cryptoxanthin is a xanthophyll, related in structure to β-carotene with only an additional
hydroxyl group. Because it also contains an ionone group, it can be converted to retinol to allow
provitamin A activity in humans. β-cryptoxanthin has exhibited protective effects against free radical
damage in cell culture, and stimulation of DNA repair. Results of studies of blood β-cryptoxanthin and
prostate cancer have been contradictory. Some studies indicate a decreased risk with higher blood
levels39, 40, while others show an increase in incidence41, 42. The top dietary sources of β-cryptoxanthin are
from fruits – tangerines, mangoes oranges and peaches, though spearmint and cilantro (coriander) also
contain high levels. Table 2 contains the top food sources in descending order.
2.6. Lutein
Lutein is a xanthophyll with no provitamin A activity. In plants it modulates light energy, and in
humans can act as an antioxidant for blue light absorption in the eye. In cell studies lutein demonstrated
selective inhibition of malignant prostate cancer cells (AT3) over their benign counterparts (DTE)43. 42%
of cancerous cells were inhibited after 4 days of culture in 2.0μM of lutein. The most substantial health
benefits of lutein however are in the eyes, where higher intakes44 and supplementation45 have been
shown to improve ocular condition. Although there are no recommended dietary intake guidelines for
lutein, positive effects (in the context of decreased risk of age-related macular degeneration) have been
seen at dietary intake levels of 6-10mg/day46.
Dietary sources of lutein & zeaxanthin are shown in Table 3, with high contents found in many
green leafy vegetables, cornmeal, beans, oranges and kiwi fruit. Lutein is approved for use as an additive
in the EU (E number E161b)47, and is commonly used in chicken feed to improve the colour of egg yolks,
and chicken skin and fat. Rohrmann et al48 showed that, compared to low intakes (0.2 servings/day),
those who consumed 1.4 servings of lutein rich food (cooked/raw spinach, kale, broccoli, Brussels
sprouts, celery, peas, and yellow squash) per day had a decreased risk of incident BPH (OR = 0.83; 95% CI
= 0.75-0.92; p value for trend = 0.0004).
Figure 4 - Differences in
Structure of Lutein and
Zeaxanthin, Adapted from
Abdel-Aal et al49
Page | 16
2.7. Zeaxanthin
Zeaxanthin is a xanthophyll, and it is one of the most common carotenoid alcohols found in
nature. Like lutein, zeaxanthin regulates light energy in plants, and is found at high concentrations in the
retina of human eyes. Higher intakes are associated with a reduced risk of age-related macular
degeneration, geographic atrophy, and large or extensive intermediate drusen44. Zeaxanthin has been
found to induce apoptosis in neuroblastoma cells, without inhibiting lipoxygenase activity or damaging
healthy cells50. An inverse association between plasma concentrations and prostate cancer was also
observed (OR = 0.22; 95% CI = 0.06–0.83; P for trend, 0.0028) when comparing highest with lowest
quartiles51.
The name “zeaxanthin” is derived from Zea mays, the trinomial term for maize/corn, and
xanthos, the Greek word for "yellow". It is the pigment that gives many foods their characteristic colours,
including paprika, saffron, corn, and egg yolks. Zeaxanthin contents combined with lutein are shown in
Table 3, but top sources of zeaxanthin alone are corn, Japanese persimmons, cornmeal, spinach, turnip
greens, collards, lettuce (cos/romaine), kale, tomatoes, tangerines, and oranges52.
Lutein and zeaxanthin share many functions and characteristics. This is because they are isomers,
differing only in the location of the double bond in one of the end rings (Figure 4). Due to their structural
similarities, many studies examine lutein and zeaxanthin together rather than individually. No studies
were found to examine dietary intake of zeaxanthin alone, and for this reason this project analyses lutein
individually, and the combination of lutein and zeaxanthin
*
.
*
“lutein & zeaxanthin” will from this point be used to describe the combined intake of the two
carotenoids – “lutein and zeaxanthin combined”
Page | 17
Table 2 – Top Food Sources of β-Cryptoxanthin,
taken from USDA Food Composition Database35
Description
µg/100 g
Spices, pepper, red or cayenne
6,252
Spices, paprika
6,186
Spices, chili powder
3,490
Squash, winter, butternut, raw
3,471
Squash, winter, butternut, cooked,
baked, without salt
3,116
Squash, winter, butternut, cooked,
baked, with salt
3,116
Tangerine juice, frozen concentrate,
sweetened, undiluted
2,767
Squash, winter, butternut, frozen,
unprepared
1,564
Persimmons, Japanese, raw
1,447
Squash, winter, Hubbard, cooked,
boiled, mashed, without salt
1,119
Peppers, hot chilli, sun-dried
1,103
Tangerines, (mandarin oranges),
canned, juice pack, drained
775
Papayas, raw
589
Tangerines, (mandarin oranges),
canned, juice pack
503
Tangerines, (mandarin oranges),
canned, light syrup pack
496
Peppers, hot chili, red, canned,
excluding seeds, solids and liquids
495
Peppers, sweet, red, raw
490
Rose Hips, wild (Northern Plains
Indians)
483
Peppers, sweet, red, cooked, boiled,
drained, without salt
460
Peppers, sweet, red, cooked, boiled,
drained, with salt
460
Peaches, dried, sulphured, uncooked
444
Tangerines, (mandarin oranges), raw
407
Peppers, sweet, red, frozen,
chopped, unprepared
380
Tamales, masa and pork filling (Hopi)
342
Soup, cream of vegetable, dry,
powder
334
Table 3 – Top Food Sources of Lutein &
Zeaxanthin, taken from USDA Food Composition
Database35. Cruciferous vegetables are in bold.
Description
µg/100 g
Kale, frozen, cooked, boiled,
drained, without salt
19,697
Kale, frozen, cooked, boiled,
drained, with salt
19,697
Spices, paprika
18,944
Kale, cooked, boiled, drained,
without salt
18,246
Kale, cooked, boiled, drained, with
salt
18,246
Spinach, frozen, chopped or leaf,
cooked, boiled, drained, without
salt
15,690
Spinach, frozen, chopped or leaf,
cooked, boiled, drained, with salt
15,690
Sweet potato leaves, raw
14,720
Dandelion greens, raw
13,610
Spices, pepper, red or cayenne
13,157
Turnip greens, raw
12,825
Spinach, frozen, chopped or leaf,
unprepared
12,651
Cress, garden, raw
12,500
Spinach, raw
12,198
Turnip greens, frozen, cooked,
boiled, drained, without salt
11,915
Turnip greens, frozen, cooked,
boiled, drained, with salt
11,915
Sweet potato leaves, cooked,
steamed, without salt
11,449
Sweet potato leaves, cooked,
steamed, with salt
11,449
Spinach, cooked, boiled, drained,
without salt
11,308
Spinach, cooked, boiled, drained,
with salt
11,308
Chard, Swiss, cooked, boiled,
drained, without salt
11,015
Chard, Swiss, cooked, boiled,
drained, with salt
11,015
Chard, Swiss, raw
11,000
Collards, frozen, chopped, cooked,
boiled, drained, without salt
10,898
Collards, frozen, chopped, cooked,
boiled, drained, with salt
10,898
Page | 18
2.8. Blood Carotenoid Levels
Although these carotenoids have exhibited anti-cancer effects, their influence on prostate cancer
incidence has not been examined to the same degree as lycopene or β-carotene. A meta-analysis of
blood carotenoids and prostate cancer risk was carried out in 200753. α-carotene, β-cryptoxanthin, and
lutein all had slightly reduced pooled relative risks – 0.97 (95% CI = 0.81-1.16), 0.96 (95% CI = 0.80-1.14)
and 0.94 (95% CI = 0.79-1.13), respectively. An increased relative risk was found for blood zeaxanthin
1.20 (0.92, 1.56), but like the other results it failed to reach statistical significance.
2.9. Objective
These four carotenoids have not been the main focus of any investigations of prostate cancer
epidemiology, and to our knowledge this is the first time that dietary intake of carotenoids has been
examined in a meta-analysis. The aim of this project was to complete a meta-analysis of dietary intake of
four carotenoids – α-carotene, β-cryptoxanthin, lutein, and lutein & zeaxanthin – to determine their role
in prostate cancer incidence.
Page | 19
3. Materials and Methods
3.1. Search Strategy/Identification of Literature
Between January 13th and February 6th an initial literature review of studies of nutritional
factors and prostate cancer (Appendix 2) was performed. Subsequently, it was decided that the focus of
this degree project should be dietary intake of carotenoids and risk of prostate cancer. All searches were
completed by a single investigator (EL), with consultation from supervisors (KB, JP, AS).
A comprehensive, systematic literature search for relevant studies was completed using
electronic databases. The primary database used was PubMed (MEDLINE), and the search comprised all
studies published up to February 2014. The Medical Subject Heading (MeSH) terms used in the search
were “Prostatic Neoplasms” AND “Carotenoids”. Studies were briefly evaluated based on their titles and
abstracts. In addition further studies were identified by reviewing the references cited in relevant articles.
The results of the search are summarised in the PRISMA54 Flow Diagram in Figure 5.
Where only abstracts were available the KI Library tool “reSEARCH” was used to locate full texts,
and to locate texts found via grey referencing. The “reSEARCH” tool amalgamates content from a
number of different sources, including PubMed, CINAHL, Cochrane Library, EMBASE, Google Scholar,
MEDLINE, OVID, and Web of Science.
3.2. Inclusion & Exclusion Criteria
Studies whose abstracts were deemed sufficient were considered for further review. Included studies
needed to meet the following criteria:
I. Contain a measure of dietary intake of carotenoids (α-carotene, β-cryptoxanthin, lutein, or lutein &
zeaxanthin)
II. Case-control, nested case-control (NCC), cohort, or case-cohort studies on human populations
III. Published as an original article
IV. Published in English and with the full text available
V. Contain an appropriate point estimate – odds ratio (OR), relative risk, rate ratio, hazard ratio (HR) –
and report 95% CIs
Studies not meeting these criteria were excluded from the analysis. Review articles and dietary
intervention trials (RCTSs)
†
were not considered for analysis. In studies with overlapping populations only
the most recent study with the largest sample size was considered. Details of the case-control and cohort
studies included in the analysis are provided in Tables 4 and 5, respectively.
†
RCT = randomised controlled trial
Page | 20
3.3. Data Extraction & Evaluation
A total of sixteen studies met the criteria for analysis. Data from these studies was extracted and
compiled into tables designed by the investigator. Where a study provided separate adjusted point
estimates for different carotenoids – α-carotene, β-cryptoxanthin, lutein, or lutein & zeaxanthin – they
were treated as independent studies.
The data that were extracted comprised the name of the first author, year of publication,
location of study, study design, sample size, age range of participants, adjusted point estimates for
highest versus lowest dietary intakes of carotenoids and corresponding 95% CIs, and adjusted covariates.
Further details of dietary assessment methods and exposure assessments were also noted (Appendices 3
and 4, respectively). The extracted data were used to manually examine study quality and also to assess
the heterogeneity of studies. These data were reviewed and approved by supervisors.
Figure 5 – PRISMA54 Flow diagram demonstrating the process of study selection
Page | 21
3.4. Statistical Analysis
For statistical purposes prostate cancer is considered a rare disease, thus an OR can be assumed
to be approximately the same as a relative risk/risk ratio. However, this does not equate or approximate
the rate ratio or HR, as summarised in the equation below:
Odds ratio ≈ (relative risk = risk ratio) ≠ (rate ratio = hazard ratio)
Studies reporting the use of Cox regression to calculate point estimate were taken as a report of HR.
These studies were analysed separately to studies using OR as their point estimates to avoid confounding
of results. Adjusted point estimates (OR or HR) and corresponding 95% CIs were pooled for highest versus
lowest dietary intakes of α-carotene, β-cryptoxanthin, lutein, or lutein & zeaxanthin, respectively.
A custom Do-File was created (Appendix 5) for statistical analysis of each carotenoid for pooled
OR and HR respectively. Publication bias was also assessed mathematically and graphically. Log ORs/HRs
and the log of the upper limit of the 95% CI for each study in the relative stratified analyses were
generated to calculate standard errors. The meta-analysis reported results in log form, but these were
converted to the natural form as shown in Appendix 6. Statistical significance was considered when
p<0.05 for pooled point estimates and when I2 >50% for heterogeneity. Publication bias was considered
present when p<0.05. All statistical analyses were carried out with Stata Statistical Software, version 13.1
(StatCorp LP, College Station, TX, USA).
Page | 22
4. Results
4.1. Study Characteristics
A total of sixteen studies satisfied the criteria for inclusion in this meta-analysis (Tables 4 and 5).
Many of the studies overlapped in their measure of different carotenoids; fifteen provided a measure of
α-carotene intake, thirteen for β-cryptoxanthin, seven for lutein, and eight for lutein & zeaxanthin. All
studies included in this analysis compared ORs or HRs of prostate cancer incidence in the highest
compared to the lowest intakes of the respective carotenoids. However these are merely qualitative
descriptors, as quantification of exposure assessments varied greatly between included studies (see
Appendix 4). Publication dates of the included studies ranged from December 1995 to February 2014.
Study design also varied widely among included studies: there were nine case-control (five
population-based and four hospital-based), four cohort, two case-cohort, and one nested case-control
study (NCC). The NCC and one of the cohort studies were conducted within the same study population,
but were both included due to differences in their designs and sub-cohort selection. All studies adjusted
for age, and he most common adjustments among included studies were for total energy intake, body
mass index (BMI), family history of prostate cancer, education, race/ethnicity, location (within respective
study population), smoking, and physical activity. The majority of studies were completed in Western
countries; ten from North America, two from Europe and one from Australia. Two were from Asian
populations and one was from South America.
4.2. Dietary Assessment
A summary of dietary assessment methods is shown in Appendix 3. All but one study61 used food
frequency questionnaires (FFQs) to assess diets of participants. In eight of the sixteen studies (including
all 6 cohort/case-cohort studies) these FFQs were completed via interview and the remaining eight were
self-administered. Nine of the questionnaires had been previously validated, and one had been
developed specifically for use in epidemiological studies of the local population. The number of
food/beverage items in these FFQs ranged from 35 to 180, and were divided into various groupings based
on each study’s individual design.
Different methods were used by each study to measure portion size and asses consumption
frequency of foods. The reference timeframe for the usual dietary intake was set at a minimum of 12
months prior to the date of assessment. United States Department of Agriculture (USDA)35 sources were
most commonly used to calculate nutrient intake estimates, with some studies using native food
composition databases. Six studies also assessed vitamin and mineral supplementation use among
participants.
Page | 23
Table 4 – Characteristics of Case-Control Studies of Dietary Intake of Carotenoids Included in Meta-analysis
Reference
(Year)
Design
n
Cases
n
Controls
Location
Age Range
(Mean)
Dietary Assessment Method
PCa Definition
Carotenoids
Investigated
Rohrmann
(2007)48
NCC
6092
18373
(HPFS)
USA
40-75
Validated, semi-quantitative
FFQ, 131 items, self-
administered
BPH – several different
criteria
α-C, β-Cr, L&Z
McCann
(2005)55
PCC
433
538
New York
not reported
FFQ, 172 items, interview
Primary histologically
confirmed prostate cancer
α-C, β-Cr, L
Bosetti
(2004)56
HCC
1294
1451
Italy
46–74
Validated FFQ, 78 food
beverages & recipes, interview
Histologically confirmed
carcinoma of the prostate
α-C, β-Cr, L&Z
Hodge
(2004)57
PCC
858
905
Melbourne, Sydney
& Perth, Australia
<70
FFQ, 121 items, interview
Histologically confirmed
prostate cancer, Gleason
score ≥5
α-C, β-Cr, L&Z
Jian
(2004)58
HCC
130
274
Hangzhou, SE China
(cases 72.7,
controls 71.4)
Validated, adapted FFQ, 130
items, interview
Histologically confirmed
adenocarcinoma of the
prostate
α-C, β-Cr, L&Z
Lu (2001)51
HCC
65
132
New York
(cases 59.98,
controls 41.9)
NCI HHHQ short dietary
questionnaire, interview
Pathologically confirmed
diagnosis of prostate
adenocarcinoma
α-C, β-Cr, L
Cohen
(2000)59
PCC
152
145
King County, WA,
USA
40-64
FFQ, 98 items, self-administered
Histologically confirmed
prostate cancer
α-C, β-Cr, L&Z
Deneo-
Pellegrini
(1999)60
HCC
175
233
Uruguay
40-89
FFQ, 64 items, interview
Histologically verified
prostatic adenocarcinomas
α-C, L
Jain
(1999)61
PCC
617
636
Ontario, Quebec, &
British Colombia,
Canada
(69.8 cases,
69.9 controls)
validated quantitative diet
history, interview
Recent, histologically
confirmed diagnosis of
adenocarcinoma of the
prostate
α-C, β-Cr, L
Meyer
(1997)62
PCC
215
593
Quebec City,
Canada
≥45
Validated FFQ, 143 items,
interview
Preclinical prostate cancer –
histologically or screen
detected
α-C, L
NCC = nested case-control, PCC = population case-control, HCC = hospital case control, HPFS = Health Professionals Follow-Up Study, FFQ = food frequency
questionnaire, NCI = National Cancer Institute (http://www.cancer.gov/), HHHQ = Health Habits and History Questionnaire, α-C = α-carotene, β-Cr = β-cryptoxanthin, L
= lutein, L&Z = lutein & zeaxanthin
Page | 24
Table 5 – Characteristics of Cohort/Case-Cohort Studies of Dietary Intake of Carotenoids Included in Meta-Analysis
Reference
(Year)
n
Cases
n
Cohort
Cohort
Name
Location
Age Range
(Mean)a
Mean Follow-
Up (Years)
Dietary Assessment Method
PCa Definition
Carotenoids
Investigated
Umesawa
(2014)63
143
15,471
JACC
Japan
40-79
(16 – median)
Validated FFQ, 35 items, self-
administered
Incident PCa
α-C
Agalliub
(2011)64
661
1,864
sub
cohort
CSDLH
Canada
(70 – age at
diagnosis)
4.3 (cases), 7.7
(controls)
Validated, adapted FFQ, 166 items,
self-administered
Incident prostate
cancer
β-Cr, L&Z
Kirsh
(2006)65
1,338
29,361
PLCO
USA
55–74
4.2 (max 8)
FFQ, 137 items, self-administered
Prostate cancer
diagnosis
α-C, β-Cr, L&Z
Stram
(2006)66
3,922
78,564
MEC
USA
45-75
7
FFQ, 180 items, self-administered
Incident prostate
cancer
α-C, β-Cr, L
Schuurmanb
(2002)67
642
1,525
NLCS
Holland
55-69
6.3
Validated, semi-quantitative FFQ, 150
items, self-administered
Incident prostate
carcinoma
α-C, β-Cr, L&Z
Giovannucci
(1995)68
812
47,894
HPFS
USA
40-75
6
Validated FFQ, 131 items, self-
administered
Adenocarcinoma of
the prostate
α-C, β-Cr, L
a = at baseline of enrolment into cohort, b = case-cohort study design, JACC = Japan Collaborative Cohort Study, CSDLH = Canadian Study of Diet, Lifestyle and Health,
PLCO = Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial, MEC = Multi-ethnic Cohort Study, NLCS = Netherlands Cohort Study, HPFS = Health Professionals
Follow-Up Study, FFQ = food frequency questionnaire, α-C = α-carotene, β-Cr = β-cryptoxanthin, L = lutein, L&Z = lutein & zeaxanthin
Page | 25
4.3. Dietary Carotenoid Intake and Prostate Cancer Risk
Results of the meta-analyses of four dietary carotenoids – α-carotene, β-cryptoxanthin, lutein,
and lutein & zeaxanthin – are summarised in Table 6. Results of statistically significant analyses are
displayed as Forrest plots in Figures 6 and 8-10. Full details of analysis of each of the four carotenoids,
including Forrest and Funnel Plots can be found in Appendix 6.
Table 6 – Summary of Results of Meta-analysis of Four Carotenoids and Risk of Prostate Cancer
Carotenoid
Point
Estimate
# Studies
Pooled OR/HR
P value
I2 (%)
P
Begg’s
P
Egger’s
α-Carotene
OR
10
0.92 (0.84-1.00)
0.04
6.3
0.15
0.06
HR
5
0.94 (0.87-1.02)
0.15
0.0
0.22
0.62
β-Cryptoxanthin
OR
8
0.91 (0.83-0.99)
0.03
76.9
0.71
0.67
HR
5
0.98 (0.91-1.06)
0.65
40.3
0.31
0.29
Lutein
OR
5
0.76 (0.60-0.97)
0.03
0.0
0.46
0.28
HR
2
1.00 (0.91-1.10)
0.97
0.0
1.00
-
Lutein &
Zeaxanthin
OR
5
0.82 (0.75-0.89)
0.00
81.9
0.22
0.29
HR
3
0.95 (0.82-1.09)
0.96
0.0
1.00
0.80
Statistically significant results in bold
4.4. α-Carotene
Fifteen studies of dietary intake of α-carotene were included in this analysis; ten case-
control/NCC, and five cohort/case-cohort. A reduced risk of prostate cancer incidence was identified for
α-carotene in both study categories. However, only the results for the case-control/NCC analysis reached
statistical significance. Analysis showed a pooled OR of 0.92 (95% CI = 0.84-1.00, p = 0.04) for higher α-
carotene intakes compared to lower (Figure 6). No significant heterogeneity was detected between
studies (6.3%), and tests for publication bias reached only borderline significance for case-control/nested
case-control studies (p value
for Begg’s test = 0.15; p value
for Egger’s test 0.06; Figure
7).
Figure 6 - Forrest Plot for
Meta-analysis of Case-
control/NCC Studies of α-
Carotene and Risk of Prostate
Cancer
Page | 26
Figure 7- Funnel Plot for Test of
Publication Bias in Case-
control/NCC Studies of α-Carotene
and Prostate Cancer
4.5. β-Cryptoxanthin
Thirteen studies included a measure of dietary intake of β-cryptoxanthin. A reduced risk of
prostate cancer was recorded in both categories of study design. However, as with α-carotene, the
analysis of five cohort/case-cohort studies did not reach statistical significance (pooled HR = 0.98; 95% CI
= 0.91-1.06). A statistically significant reduced risk of prostate cancer was recorded in the analysis of the
eight case-control/NCC studies (pooled OR = 0.91; 95% CI = 0.83-0.99; p=0.03; Figure 8). There was
significant heterogeneity between studies in this analysis (I2 = 76.9%, p = 0.000), and possible reasons for
this are discussed in section 5.4. Publication bias was not detected in neither case-control/NCC (p Begg’s
=0.71; p Egger’s = 0.67) nor cohort/case-cohort studies (p Begg’s =0.31; p Egger’s = 0.29).
Figure 8 - Forrest Plot for Meta-analysis of case-control/NCC Studies of β-Cryptoxanthin and Prostate Cancer
0
.2
.4
.6
s.e. of logOR
-1.5 -1 -.5 0 .5 1
logOR
Funnel plot with pseudo 95% confidence limits
Page | 27
Figure 9 - Forrest Plot for Meta-analysis of Case-control/NCC Studies of Lutein and Risk of Prostate
Cancer
4.6. Lutein
Only seven studies included a measure of dietary intake of lutein – five case-control and two
cohort studies. In the cohort study analysis, no association was found for the highest category of lutein
intake (pooled HR = 1.00; 95% CI =0.91-1.10; p = 0.97), but this result is limited due to the small number
of studies included. There was a statistically significantly reduced risk of prostate cancer found in the
analysis of case-control studies (pooled OR = 0.76; 95% CI = 0.60-0.97; p = 0.03; Figure 9). No
heterogeneity was detected in either the case-control (0.0%) or cohort studies (0.0%), and no publication
bias was detected among case-control studies in either of Begg’s (p = 0.46) or Egger’s tests (p = 0.28).
Assessment of publication bias in cohort studies of lutein was not possible due to the limited number of
studies available for analysis.
4.7. Lutein & Zeaxanthin
Eight studies examined the intake of lutein and zeaxanthin combined – five case-control/NCC and
three cohort/case-cohort. Similar to the other cohort/case-cohort analyses the pooled HR did not reach
statistical significance (0.95; 95% CI = 0.82-1.09; p = 0.96), and no heterogeneity was observed (0.0%). In
the case-control/NCC analysis, a statistically significant reduced risk of prostate cancer was associated
with the highest intakes of lutein & zeaxanthin (pooled OR 0.82; 95% CI = 0.75-0.89; p = 0.000; Figure 10).
However, there was significant heterogeneity (81.9%, p = 0.000), which is discussed along with β-
cryptoxanthin in section 5.4. No significant publication bias was detected among case-control/NCC (p
Page | 28
Figure 10 - Forrest plot for Meta-analysis of case-control/nested case-control studies of lutein &
zeaxanthin and Risk of Prostate Cancer
value for Begg’s test = 0.22; p value for Egger’s test 0.29), or cohort/case-cohort studies. Results of the
publication bias analyses among cohort/case-cohort studies is limited due to the low number of studies
available for analysis (p value for Begg’s test = 1.00; p value for Egger’s test 0.80).
Overall these results demonstrate an inverse association of increasing intakes of α-carotene, β-
cryptoxanthin, lutein, and lutein & zeaxanthin.
Page | 29
5. Discussion
To our knowledge, this is the first time that these four carotenoids – α-carotene, β-
cryptoxanthin, lutein, and zeaxanthin – have been the focus of a meta-analysis of dietary intakes and
prostate cancer incidence, despite there being ample data available from multiple epidemiological
investigations. The majority of studies took place in Western populations, which may reflect their higher
incidences of prostate cancer (Figure 2). A reduced risk of prostate cancer was found for higher intakes of
all four carotenoids, however only the results for case-control/NCC studies reached statistical significance
(see sections 4.3-4.7). Lutein showed the strongest protective effect, with a reduced incidence of 24%
found among those with higher intakes in case-control studies. Combined intakes of lutein & zeaxanthin
also reduced incidence at higher intakes, as did α-carotene and β-cryptoxanthin. However the results of
these analyses are disputed in section 5.5 below.
5.1. Measures of Association
Studies were categorized based on their measure of association, OR or HR. Studies reporting a
measure of risk ratio or relative risk were reviewed to see if there methods were statistically sound.
Three studies used “relative risk” as their outcome measure but were reassigned HRs due to their
statistical reasoning. Two of these65, 66 were because they used Cox regression and, another68 for using
rates in their calculation and proportional hazards regression. Schuurman et al67 used “rate ratio” for
their outcome measure and assumed exponentially distributed survival times, which was judged to be a
HR calculation. Two studies reported an outright HR measurement and were accepted as such63, 64. All
studies reporting HR were either cohort or case-cohort, whereas all studies using OR as their outcome
measure were case-control or NCC. Due to these differences in study design and point estimate
measures, is was not possible to measure the cumulative effect among all studies for each carotenoid.
5.2. Study Design
Rohrmann et al48 carried out a nested case-control study in their investigation, using data from
the Health Professionals Follow-Up Study. An NCC study is one where subjects are sampled from an
already assembled epidemiological cohort study, in which the sampling depends on disease status69.
Case-cohort designs are similar to nested case-control studies, except that the controls are randomly
selected from the full cohort without matching. Case-cohort studies do however allow for the evaluation
of multiple disease endpoints70, which was ideal for Agalliu et al64, as they examined individual pro- and
anti-oxidants as well as their cumulative influence on prostate cancer risk. Similarly Schuurman et al67
assessed the intakes of certain nutrients and incidence of prostate cancer among drinkers and non-
Page | 30
drinkers. Characteristics of case-control/NCC and cohort/case-cohort studies are shown in Tables 4 and 5,
respectively.
Significantly reduced incidences of prostate cancer were found for each of α-carotene, β-
cryptoxanthin, lutein, and lutein & zeaxanthin in all case-control/nested case-control studies. Though the
results for cohort/case-cohort studies show reduced risks (except for lutein, which showed no
association; pooled HR = 1.00; 95% CI = 0.91-1.10), they did not reach statistical significance. This pattern
has been seen in previous meta-analyses of dietary intakes and prostate cancer risk19, 71. A possible
reason for this could be the high number of hospital-based case-control studies included in this analysis,
as these are more vulnerable to selection and Berkson’s bias72 than population case-control studies and
cohort studies.
Case-control studies are generally considered less consistent, and rank lower on the hierarchy of
evidence73. This could be because case-control studies are vulnerable to oversampling, in that the
number of cases (and matched controls) may not be representative of disease rates in the entire
population. Location of studies may be another reason for this, as studies carried out in low-risk
populations (e.g. Asia) may not contain a sufficient number of cases to be representative of total
incidence worldwide. The largest cohort study by Stram et al66 came from the Multi-ethnic Cohort
Study74, and contained 3,922 cases and 78,564 controls. A study like this would be more representative of
prostate cancer incidence in the total population than a case-control study. However, the purpose of this
analysis was to determine the influence of dietary carotenoids on prostate cancer risk, and case-control
studies are an invaluable source of information in epidemiological investigations such as this one.
5.3. Dietary Assessment
Despite the high quality of dietary assessment methods among studies in this analysis, there was
a lot of inconsistency (see Appendix 3). All but one study61 used food frequency questionnaires (FFQs) to
assess participants’ diets, and although these FFQs were tailored to the populations being scrutinized,
there is little opportunity to make comparisons between studies. Umesawa et al63 only included 35 foods
in their FFQ, and gave no information about how the nutrient and carotenoid contents were calculated.
Although the FFQ used for that study had been previously validated, this does not sufficiently reflect the
wide variety of foods containing high levels of different carotenoids. Half of the studies included
measured between 121 and 166 food items in their FFQs. In an attempt to reduce recall bias75, most
studies asked participants to estimate their consumption over the past year. This would also account for
any changes in diet that may have occurred following prostate cancer diagnosis76.
Ten studies utilized the UDSA Nutrient Database to calculate the carotenoid contents of the food
they investigated. Two studies used native food composition databases (Italian56 and Dutch67), and
Page | 31
another64 adapted the UDSA data to reflect local food availability and fortification laws. Adaptation of
nutrient estimates or use of local estimates is favourable, as soil quality and nutrient content vary widely
throughout the world, and the use of databases from other countries may result in incorrect calculations
of nutrient composition of foods. Two studies used composition data from Mangels et al77, and a further
two utilised the Nutrition Data System developed by the University of Minnesota78. Six studies also
examined supplement use, which could have contributed to higher carotenoid intakes. Supplementation
is further discussed in section 5.9.
There were also significant differences in the techniques used to calculate consumption
frequency and portion size. This may in some part explain the disparities in the categories of intake used
for analysis (see Appendix 6). Point estimates (OR or HR) were taken for the highest versus lowest level of
intake, be they quintiles, quartiles, mean, or median values. Stram et al66 defined quintile boundaries
based on micrograms per 1,000 kilocalories, preventing comparison to other studies which estimated
daily intakes, and Meyer et al62 did not report the quartile boundaries used in their study. Using lutein as
an example, the quartile boundaries set by McCann et al55 are particularly high. Their first/lowest quartile
of intake (≤3029µg/d) was higher than the fourth /highest quartile used by Jain et al61 (>2684 µg/d).
These exceptionally high intake measures could account for the significant trend found for increased
lutein intake reported my McCann et al (p = 0.01). In all cases, disparities in categorizing intake levels
contributed to heterogeneity among studies.
5.4. Heterogeneity
Further heterogeneity between studies can be attributed to the adjustments used in point
estimate calculations (see Appendix 4). All sixteen studies adjusted for age, an already established risk
factor for cancer79. The next most common adjustments were for BMI (11/16), energy intake, family
history of prostate cancer (10/16), education, location (6/16), smoking, race/ethnicity (5/16), fat intake,
alcohol intake (4/16), physical activity, socioeconomic status (3/16), and marital status (2/16). The
number of adjustments applied in each study also varied, with most studies adjusting for four to fourteen
covariates. Giovannucci et al68 only adjusted for age and energy intake, whereas Jain et al61 adjusted for a
total of 23 covariates, many of them log-converted amounts for other dietary factors.
The sample sizes also varied between the studies included in the analysis. Among cohort studies,
the largest cohort study (Stram et al66, The Multi-ethnic Cohort Study74) had over five times more
participants than the smallest one (Umesawa et al63, Japan Collaborative Cohort80). However, there were
only 143 cases of prostate cancer in the JACC study, compared to 3,922 in the MEC one. This rate is over
24 times higher, though probably attributable to the lower incidence rates in Japan1. In case-control
Page | 32
studies, the smallest one had just 65 cases and 132 controls51, while the largest one had 1294 cases and
1451 controls56.
Rates for prostate cancer are low among men below the age of 45 (9.2 per 100,000 for men aged
40-44 years), but increase to 984.8 per 100,000 in men aged 70-74 years79. The majority of studies
enrolled men aged between 45 and 75 and only one study included men below the age of forty51. Only
one study allowed men over the age of eighty60, though McCann et al55 did not report the ages of
participants.
5.5. Exclusion of Jian et al58
Although there was a reduced risk of prostate cancer found in the case-control/NCC analyses of
β-cryptoxanthin (pooled OR = 0.91) and lutein & zeaxanthin (pooled OR 0.82), there was substantial
heterogeneity observed in these analyses. The results of the I2 tests showed 76.9% and 81.9%
heterogeneity among studies for β-cryptoxanthin and lutein & zeaxanthin, respectively. Possible sources
for this heterogeneity have been discussed above, but manual examination of the ORs for the individual
studies involved in these analyses led us believe that the study by Jian et al58 could be a potential outlier.
To test this hypothesis we repeated the meta-analysis of case-control/NCC studies for β-
cryptoxanthin and lutein & zeaxanthin excluding the results from Jian et al (Appendix 7). For β-
cryptoxanthin, heterogeneity was reduced to 47.7%, though the reduced risk of prostate cancer did lose
statistical significance (pooled OR 0.93; 95% CI = 0.85-1.01; p = 0.08). In the case of lutein & zeaxanthin,
exclusion of the study by Jian et al completely eliminated heterogeneity (0.0%). The protective effect of
higher intakes was slightly diminished, but maintained statistical significance (pooled OR 0.83; 95% CI =
0.76-0.90; p = 0.000). There was no detectible publication bias found for neither β-cryptoxanthin nor
lutein & zeaxanthin, in repeat Begg’s or Egger’s tests (results not shown).
Though no significant heterogeneity was found among case-control/nested case-control studies
of α-carotene (6.3%), we repeated the analysis excluding Jian et al. Heterogeneity was reduced to 0.0%,
but like β-cryptoxanthin, the results lost statistical significance (pooled OR 0.93; 95% CI =0.85-1.01; p =
0.08). We can conclude that the results reported by Jian et al are strongly influential to the analysis of α-
carotene and β-cryptoxanthin. Results for lutein & zeaxanthin however remained robust, and their
influence in prostate cancer should be further examined in future studies.
Upon initial manual examination of Jian et al, no significant differences between other studies
were detected. Possible reasons for the exceptionally low ORs found in this study include the large
reference recall period for dietary assessment interview (5 years before diagnosis/interview), or the fact
that 65.3% of control were recruited from the inpatient urology department of the hospitals involved.
This is the only case-control study carried out in Asia, the region with the lowest rates in the world, and
Page | 33
this is reflected in the relatively small sample size. Although the FFQ used in this study was adapted from
questionnaires used in four previous studies, it had been previously validated in native Chinese
populations. The authors assume that participants in the study were representative of the Zhejiang
population, and the FFQ used contained traditional Chinese units of measurement (“Liang”). It is possible
that the low rates of prostate cancer in the region examined by Jian et al contributed to the exceptionally
low ORs found for higher carotenoid intakes.
5.6. Prostate Cancer Definition
There was some variance in the definition of prostate cancer between studies. Though not all
studies specify, we can assume that “prostate cancer” refers to adenocarcinoma of the prostate. In many
of the studies, the prostate cancer needed to be histologically or pathologically confirmed before a
participant could be considered a case. In the cohort studies, subjects with prostate cancer at enrolment
were excluded from their investigations. Incident prostate cancer was used to detect cases in these
studies, usually through linkages with local cancer registries.
We included two studies that used benign prostatic hyperplasia (BPH) as their inclusion criterion.
Over 90% of men aged 85 show histological evidence of BPH, and approximately one in four men will
require medical care for the condition by age 8081. Though the link between BPH and prostate cancer has
been disputed82, both conditions have high prevalence worldwide and share many pathophysiological
properties. Finasteride is a 5α-reductase inhibitor used to treat BPH83, and this drug was also shown to
reduce overall risk of prostate cancer by 30% in the Prostate Cancer Prevention Trial84. Lycopene has
been shown to reduce BPH progression85, and so the effects of other carotenoids on the condition
warrants further investigation.
Two of the studies used BPH as their inclusion criterion. Meyer et al62 assessed two groups; one
was men hospitalised for transurethral prostatectomy (TURP) where prostate cancer was discovered in
resected tissue, and the other was men who took part in a screening program and were referred for
radical treatment during the study period. These two groups were combined for the analysis of nutrient
intake, and cases are referred to as “preclinical prostate cancer” throughout the paper. Rohrmann et al48
carried out an NCC study within the Health Professionals Follow-Up Study (HPFS), which involved a
number of different follow-up assessments during the study period. They examined BPH, defined in two
different ways. Diagnosis of “total BPH” was based on a history of surgery for an enlarged prostate, high-
moderate to severe lower urinary tract symptoms (AUASI
‡
score ≥15) and use of medications (α-blockers,
finasteride) to treat BPH. Diagnosis of “incident BPH” was based reports of surgery/symptoms in follow-
ups after 1994. In this analysis we took the OR for “total BPH”.
‡
AUASI = American Urological Association Symptom Index
Page | 34
5.7. Publication Bias
We assessed whether or not publication bias was present using Begg’s and Egger’s tests and by
generating funnel plots (see Appendix 5). Publication bias is the selective publication of studies based on
favourable characteristics86, for example studies reaching statistical significance, popularity of the topic,
having a sponsor, and studies published in English. The Begg’s test assesses the presence of association
between the effect estimates and their variances87, with significant correlation indicating publication bias
is present. This test however is unreliable when the number of studies is small, so we also used Egger’s
test88 which is more specific. Egger’s test plots a regression line between precision of the studies and the
standardized effect, and measures correlation mathematically to generate a p value like Begg’s test.
No publication bias was detected in any of the stratified analyses. Tests for α-carotene reached
borderline significance in the Egger’s test (0.06) in case-control/nested case-control studies, and slight
asymmetry (caused by two outliers) was noted in the funnel plot (Figure 7). The probability of publication
bias being influential in this analysis is low, as none of the carotenoids examined here have been the
main topic of any epidemiological investigations into prostate cancer. However, because this project
specified non-English papers as an exclusion criterion, publication bias has been introduced. This was due
to time constraints, and can be amended for future studies.
5.8. Limitations of Current Analysis
Nutritional epidemiology investigations are always quite limited in their power due to the
retrospective nature of dietary recall, and the limited timeframe to complete this project means that
there are further limitations in the methodology. Firstly, all reviews, investigation procedures and data
extraction were carried out by a single investigator (EL). This introduces the possibility of bias in
assessment and recording, and ideally the literature review should have been completed in tandem with
another investigator. Secondly, publication bias was introduced as only studies published in English were
considered. Also no contact with authors was made when full texts were unavailable, which could
potentially have added to our results pool. Thirdly, there was significant heterogeneity among studies,
particularly in exposure measurements and adjusted covariates. Excluding Jian et al as an outlier did
reduce percentage heterogeneity in all categories affected, but led to the results of case-control/NCC
analyses of α-carotene and β-cryptoxanthin losing statistical significance.
5.9. Supplementation
Six of the studies included in this meta-analysis examined supplement use among participants,
which could have contributed to carotenoid intakes (see Appendix 3). A recent systematic review and
Page | 35
meta-analysis89 assessed multivitamin use and prostate cancer occurrence. Neither multivitamin
supplementation nor use of individual vitamins or minerals (vitamin E, zinc, selenium, and β-carotene)
affected the overall occurrence of prostate cancer. Mortality and incidence of high-grade/metastatic
prostate cancer were not affected either, and though there was considerable heterogeneity between the
studies, stratified analysis of high-quality studies returned similar results. Another review and meta-
analysis of RCTs90 examined the influence of supplementation with non-herbal dietary supplements and
vitamins on prostate cancer patients. They found that no evidence that dietary supplements reduced PSA
levels, though two trials using mixtures including carotenoids, lycopenes, and antioxidants (among many
others) did significantly reduce PSA levels.
Carotenoid supplementation has not been the subject of many prostate cancer prevention trials.
A Cochrane review of lycopene supplementation91 found that there was insufficient evidence to either
support or refute the use of lycopene for the prevention of prostate cancer. Further evidence is required
to determine whether carotenoid supplementation is a viable preventative mechanism for prostate
cancer incidence or progression.
5.10. Synergy
Despite a lack of evidence from RCTs about the use of carotenoids as a chemoprevention
mechanism for prostate cancer, epidemiological investigations of high carotenoid food have returned
significant results. As mentioned in section 2.3, lycopene has been extensively investigated as a possible
preventative agent for prostate cancer. However, the results of studies investigating foods with high
lycopene concentrations, such as tomatoes, have returned even more favourable results. The meta-
analysis by Chen et al32 showed greater reductions in ORs for higher intakes of tomatoes than for
lycopene intake. Similarly, the most recent meta-analysis of cruciferous vegetable intake19 demonstrated
a relative risk of 0.90 (95% CI = 0.85-0.96) for higher intakes compared to lower, and many cruciferous
vegetables are high in lutein and zeaxanthin (Table 3).
Studies involving whole foods or food groups tend to show greater risk reduction of prostate
cancer, and are more often statistically significant than studies of nutrients alone. Carrots24, tomatoes32,
and cruciferous vegetables19 have all been shown to reduce prostate cancer incidence in high consumers.
As shown in Tables 1-3, these foods contain high levels of carotenoids. It is possible that these foods
contain other biomolecules that strengthen the protective effect of carotenoids in a synergistic way.
These studies reinforce the findings made here, that higher carotenoid intakes reduce prostate cancer
risk.
Page | 36
5.11. Carotenoid Bioavailability
Tables 1-3 show the top dietary sources of α-carotene, β-cryptoxanthin, and lutein & zeaxanthin,
respectively. From these tables we can see that there is significant variance between contents based on
their preparation and cooking methods. In the case of α-carotene, carrots make up thirteen of the top 25
food sources. Dehydrated carrots, the top dietary source of α-carotene, have over 4 times the amount of
α-carotene as raw carrots. Thermal processing of tomato products is an effective way to increase their
carotenoid concentrations92. Giovannucci et al68 recorded a relative risks of 0.66 (95% CI = 0.49-0.90) for
men who consumed 2-4 servings of tomato sauce per week compared to those who had none (p for
trend 0.001). Protective effects can also be greater for different varieties of foods that reduce risk93.
Different methods of cooking or preparing foods can have a significant impact on the
bioavailability of carotenoids in foods. As they are lipid soluble, it has been hypothesised that increasing
fatty acid intake as well as carotenoids can improve their health benefits. A 2005 study94 found that
addition of avocado (in fruit or oil form) to salads and salsa enhanced absorption of α-carotene and lutein
(p < 0.01). Although saturated fat has been shown to increase prostate cancer risk at higher intakes23,
other fatty acid subgroups have been shown to decrease risk (see Appendix 1), and higher quality dietary
fats can contribute to better overall health outcomes95.
5.12. Interpersonal Differences
Epidemiological investigations and meta-analyses give generalised results on how different
factors can affect disease outcomes. There are many confounding factors in prostate cancer
development, evident by the differences in adjusted covariates between studies (see Appendix 4).
Interpersonal differences between members of a population can influence an individual’s chance of
developing a disease. Different genotypes96, ethnicities97, and lifestyle choices (see section 2.1) carry
different risks of developing prostate cancer, and thus the influence of carotenoids as a preventative tool
will be variable.
Plasma carotenoid concentration and dietary intake are correlated98, though differences in
plasma levels may be due to the different uptake rates among other tissues. One study99 found that body
fat influenced the tissue distribution of carotenoids, with significantly higher concentrations in abdominal
adipose tissue compared to the buttock and thigh. Another study100 found that lycopene and β-carotene
were also found in high concentrations in skin compared to lutein and zeaxanthin, and total carotenoids
were significantly correlated in skin and plasma. Strong associations between serum and colon
measurements of α-carotene and β-cryptoxanthin have also been observed101. Furthermore, a number of
different carotenoids have been found to be intercorrelated in prostatic tissue102, with disparities in
Page | 37
concentrations among benign and malignant samples. Although all these measures are correlated,
differences in metabolism may account for disparities between subjects.
5.13. Plant-Based Diets
There is substantial evidence to suggest that vegetarian or vegan diets improve overall health103,
104. In the context of prostate cancer, red and processed meat105 and eggs106 have been shown to
increase prostate cancer risk when consumed in higher amounts. There is also substantial evidence that
soy foods17, a common alternative to meat in vegetarian and vegan diets, have a protective effect against
prostate cancer. Carotenoids are obtained almost exclusively from plant sources, so increasing intakes of
all fruits and vegetables should decrease prostate cancer risk and improve overall morbidity. Increased
fruit and vegetable intake would also be beneficial, as intakes of specific fruits and good predictors of
certain individual plasma carotenoid levels107. Increased plant food from high carotenoid sources such as
carrots, cruciferous vegetables (kale, spinach, turnip greens, collards, etc.), tomatoes, pumpkin, squash,
tangerines/mandarins/oranges, and peppers should be incorporated into nutrition guidelines for prostate
cancer prevention.
5.14. Future Directions: Developing a Prostate Cancer Diet Score (PCDS)
Compiling results of meta-analyses of dietary factors and their influence on prostate cancer
would facilitate the creation of a dietary assessment tool. This would be in conjunction with the
introduction of a set of dietary guidelines targeting at-risk groups for prostate cancer. Dietary guidelines
as a chemopreventative measure or treatment for preclinical prostate cancer would eliminate the
burdens of over diagnosis and redundant therapy for non-fatal cases3. Adherence to dietary
modifications has been shown to be favourable108, 109, so these guidelines and PCDS would provide a
simple, cost-effective, easy-to-use treatment option for low-grade prostate cancer.
More studies of dietary investigations will need to be completed before these guidelines and
PCDS can begin to take shape. Possible candidates identified during the initial literature review for this
project include allium vegetables, carbohydrates (flour, grains, sugar, etc.), phosphorous, zinc, iron,
different fish types, and animal and plant proteins. There are sufficient data available for a meta-analysis
of β-carotene, and other carotenoids have also demonstrated anti-cancer mechanisms110. Further work is
need to determine the influence of retinol and vitamin A, particularly their association with the
provitamins in this project. Results of already established analyses would also have to be re-evaluated to
quantify ideal intakes.
Page | 38
Some studies have assessed dietary patterns and their effect on prostate cancer by creating
specified food matrices and examining overall effects of food intakes111-113. One of the studies included in
this analysis64 also developed an oxidative balance score based on the pro- or anti-oxidant effects of
different factors on prostate cancer development. Once a sufficient number of relevant dietary factors
with significant influence (be it positive or negative) on prostate cancer have been accumulated, their
influence will be assessed114 and quantified, and statistical models can be built to create the PCDS.
6. Conclusion
This meta-analysis of four dietary carotenoids has shown that there is a reduced risk of prostate
cancer among men with higher intakes of α-carotene, β-cryptoxanthin, lutein, and zeaxanthin. Increased
intakes of high carotenoid foods may be protective against prostate cancer development. Further
research and quantification of ideal intake levels is required before recommendations for a prostate
cancer diet score can be determined.
Page | 39
Appendix 1 – Results from Most Recent Meta-analyses of
Dietary Factors and Prostate Cancer Incidence
Nutrient/Food
Risk Estimate
95% CI
Most Recent
Meta-Analysis
Reference
Coffee
0.88
0.82–0.95
2014
Cao21
Green Tea
0.79
0.43-1.14
Lin71
Black Tea
0.88
0.73-1.02
Carrots
0.82
0.70-0.97
Xu24
ALA
1.08
0.90-1.29
2013
Carleton115
Fruit
1.02
0.98-1.07
Meng116
Veg
0.97
0.93-1.01
Raw Tomato
0.81
0.59-1.10
2012
Chen32
Cooked Tomato
0.85
0.69-1.06
Lycopene
0.93
0.86-1.01
Long Chain n-3 PUFAs
1.06a
0.88-1.28
Chua117
Arachadonic Acid
1.09a
0.97-1.23
DHA
0.99a
0.92-1.07
EPA
1.00a
0.92-1.08
Linoleic Acid
0.97a
0.86-1.10
Total Omega 3
0.97a
0.89-1.07
Total Omega 6
1.04a
0.95-1.13
Cruciferous Veg
0.90
0.85–0.96
Liu19
Alcohol
1.08
0.97–1.20
Rota118
Fish
0.85
0.72-1.00
Szymanski20
Egg
1.09b
0.86-1.31
Xie106
Vitamin D
0.83b
0.28-2.43
2011
Gilbert119
Processed Meat
1.05
0.99-1.12
2010
Alexander105
Red Meat
1.00
0.96-1.05
Daidzein
0.66
0.51-0.86
2009
Hwang17
Genistein
0.67
0.52-0.86
Non-Fermented Soy Food
0.75
0.62-0.89
Tofu
0.73
0.57-0.92
Total Soy Food
0.69
0.57-0.84
Soybean Milk
0.57
0.19-1.76
Dairy Products
1.11a
1.03–1.19
2008
Huncharek22
Sat Fat
1.09
0.99-1.20
2004
Dennis23
Total Fat
1.17
1.10-1.25
Statistically significant results highlighted; a = figure for cohort studies only, b = figure for case-control
studies only
Page | 40
Appendix 2 – Details of Initial Literature Review of Dietary
Factors and Prostate Cancer
Between January 13th and February 6th a preliminary literature review of dietary factors and
prostate cancer was performed using the electronic PubMed (MEDLINE) database. The search terms were
"nutrition" OR "diet" AND "prostatic neoplasms", and the search comprised all studies published up to
January 2014. Studies were briefly evaluated based on their titles and abstracts. In addition further
studies were identified by reviewing the references cited in relevant articles. Laboratory studies and
dietary intervention trials were not considered for review. The results of the search are summarised in
Figure A below.
Figure A – PRISMA54 Flow Diagram Summarising Initial Literature Review of Dietary Factors and Prostate
Cancer Incidence
Page | 41
Appendix 3 – Summary of Dietary Assessment Methods Used in Included Studies
Reference
FFQ
Validated
Adapted
# Items
Carotenoid
Estimates
Self-
Administered
Questionnaire
Interview
Timeframe
Portion Size
Frequency of
Consumption
Supplement Use
Agalliu64
yes
yes
NCI Canada
166 food items
USDA nutrient
database modified
to reflect Canadian
food availability
and nutrient
fortification laws.
yes
no
Usual intake over
past year
Usual (average)
portion size
Multivitamin and
single supplement
and mineral use - #
pills/week, # months
of use
Bosetti56
yes
yes
78 foods,
beverages &
recipes
Italian food
composition
database (Salvini et
al)
no
yes
Usual diet during
the 2 years prior to
cancer
diagnosis/hospital
admission
Weekly frequency
of consumption of
each dietary item
Cohen59
yes
99 food items,
including 12
fruit items and
21 vegetable
items
Incorporated
updated data from
the USDA on
carotenoid content
of fruits and
vegetables
yes
no
3-5 year period
preceding
reference dates
3 options for
portion size
9 options for
frequency
Deneo-Pellegrin60i
yes
no
64
Mangels et al (1993
USA)
no
yes
Past year/year prior
to onset of
symptoms
A commonly
used unit/portion
size was specified
for each food,
open-ended
responses
Responses
converted to times
per year
Page | 42
Giovannucci68
yes
yes
131 food and
beverage
items, 46 fruit,
veg & related
items
USDA sources
yes
no
Past year
Commonly used
unit or portion
size was specified
for each food
item
9 possible
responses
Brand, duration, and
frequency of
multivitamin and
individual vitamin
supplement use
Hodge57
yes
Developed
specifically for use
in Australian
epidemiological
studies
121 item FFQ -
29 groups
(some
subgroups of
others)
Version 11 of the
USDA carotenoid
database
no
yes
Jain61
no
yes
Quantitative diet
history
encompassing
1,129 unique food
items
Classified into
29 food groups
for analysis
USDA-National
Cancer Institute
carotenoid food
composition
database
no
yes
One year prior to
diagnosis/interview
date
Jian58
yes
yes
Modified from FFQs
from 4 other
sources
130 food items
USDA nutrient
database
no
yes
5 years before
diagnosis/interview
Options ranging
from 0-2
times/year to ≥2
times/day
Kirsh65
yes
137 food items
University of
Minnesota
Nutrition Data
System for
Research
yes
no
Previous year
Usual portion
size (S/M/L)
Multivitamins &
single-nutrient
supplements -
duration, frequency,
dose/day, when they
started taking them
Lu51
yes
NCI HHHQ short
FFQ
Nutrient contents
calculated using an
NCI algorithm,
based on USDA
database
no
yes
One year prior to
diagnosis/interview
date
Usual dietary
patterns, usual
portion size
Frequency of
consumption
Page | 43
McCann55
yes
no
Comparable to
FFQs used by NCI &
Harvard NHS
172
Food composition
data from USDA
no
yes
2 years prior to
interview
Included info on
portion size
Meyer62
yes
yes
Food list modified
to better reflect the
dietary habits of
the French-
speaking
population of
Quebec, and
expanded to
improve the dietary
assessment of fat,
retinol, and
carotenoid intake
143 foods or
dishes
Mangels et al 1993,
micronutrient
intake from
supplements was
computed using the
1993 Canadian
Compendium of
Pharmaceutical
Specialties
no
yes
Previous 12 months
Three-
dimensional
models were
used to
determine
portion size
Intake frequency
Intake of vitamin
and mineral
supplements, name
and amount of each
supplement,
frequency and
duration of use
Rohrmann48
yes
yes
131 item semi-
quantitative
USDA sources
yes
no
Commonly used
unit or serving
size specified for
each food item
9 possible response
categories ranging
from “never” to “6
or more times per
day”
Dose and duration of
vitamin supplement
intake
Schuurman67
yes
yes
150 items
semi-
quantitative
FFQ
Goldbohm et al
1988
yes
no
Year preceding
start of study
Habitual
consumption
Any vitamin
supplement usage
during five years
before baseline
Page | 44
Stram66
yes
180
USDA nutrient
database
yes
no
Previous year
Photographs of
representative
food items
showing three
different portion
sizes were used
to facilitate
quantification of
intakes
Frequency of
consumption
.
Umesawa63
yes
yes
35 foods
yes
no
5 responses were
possible ranging
from ‘rarely’, to
‘almost every day’
Page | 45
Appendix 4 – Exposure Assessment Boundaries and Adjusted
Covariates
4.1. α-Carotene
Reference
Adjustments
Measurement
Q1
Q2
Q3
Q4
Q5
Bosetti56
age, study centre, education, physical
activity, body mass index, family history of
prostate cancer and total calorie intake
µg, mean
values (SD)
not reported
Cohen59
fat, energy, race, age, family history of
prostate cancer, body mass index,
prostate-specific antigen tests in previous
5 years, and education
µg
<330
330-549
550-809
≥810
Deneo-
Pellegrini60
age, residence, urban/rural status,
education, family history of prostate
cancer, body mass index and total energy
intake
µg/day
≤109
110–291
292–600
601+
Giovannucci68
energy-adjusted nutrient, adjusted for age
by stratified analysis
µg
<380
380-522
523-722
723-
1339
>1339
Hodge57
state, age group, year, country of birth,
socioeconomic group, and family history
of prostate cancer
µg/day
670-
1243
1244-
1497
1498-
2125
2126+
Jain61
log total energy, vasectomy, age, ever-
smoked, marital status, study area, body
mass index, education, ever-used
multivitamin supplements in 1 yr. before
diagnosis/interview, area of study, and
log-converted amounts for grains, fruit,
vegetables, total plants, total carotenoids,
folic acid, dietary fibre, conjugated linoleic
acid, vitamin E, vitamin C, retinol, total
fat, and linoleic acid
µg/day
<839
839-
1514
1515-
2187
>2158
Jian58
age at interview, BMI, locality of
residence, education, family income,
marital status, number of children, family
history of prostate cancer, tea drinking,
total caloric intake (kcal/day) and total
fat intake (gm/day)
µg/day
<238.9
238.9-
747.5
747.6-
1786
>1786
Kirsh65
age, total energy, race, study centre,
family history of prostate cancer, body
mass index, smoking status, physical
activity, total fat intake, red meat intake,
history of diabetes, aspirin use, and
number of screening examinations during
median
µg/day
472
784
1081
1476
2317
Page | 46
the follow-up period
Lu51
age, race, education, alcohol drinking,
pack-years of smoking, family history of
prostate cancer, and total dietary caloric
intake
µg
<385.765
385.766-
699.293
699.294-
1142.31
>1142.32
McCann55
age, education, body mass index,
cigarette smoking status, total energy, veg
intake
µg/day
≤626
626-977
977-
1488
>1488
Meyer62
age, education, family history of prostate
cancer, group, dietary energy
Rohrmann48
age, race or ethnicity, cigarette smoking,
BMI, leisure-time physical activity, alcohol
consumption, energy intake, intake of
protein, and intake of polyunsaturated
fatty acids
µg/day
293
493
634
1019
2040
Schuurman67
age, family history of prostate cancer,
socioeconomic status, and alcohol from
white or fortified wine
mg/day
0,2
0,4
0,6
0,8
1,3
Stram66
age, BMI, education and family history of
prostate cancer
µg/1000kcal
≤170.8
170.9-
264.2
264.3-
382.7
382.8-
623
≥623.1
Umesawa63
age, BMI, ethanol intake, smoking status,
daily green tea intake and work schedule
median
µg/day
105
175
236
317
496
Page | 47
4.2. β-Cryptoxanthin
Reference
Adjustments
Measurement
Q1
Q2
Q3
Q4
Q5
Agalliu61
age at baseline, race,
BMI, exercise activity,
and education
µg, median values
83.3
164.5
211.3
269.1
388.2
Bosetti56
age, study centre,
education, physical
activity, body mass
index, family history of
prostate cancer and
total calorie intake
µg, mean values
(SD)
Not reported
Cohen59
fat, energy, race, age,
family history of
prostate cancer, body
mass index, prostate-
specific antigen tests in
previous 5 years, and
education
µg
<10
10-24
25-44
≥45
Giovannucci68
energy-adjusted
nutrient, adjusted for
age by stratified analysis
µg
<22
22-40
41-67
68-114
>114
Hodge57
state, age group, year,
country of birth,
socioeconomic group,
and family history of
prostate cancer
µg/day
152-272
273-415
416-657
658+
Jain61
log total energy,
vasectomy, age, ever-
smoked, marital status,
study area, body mass
index, education, ever-
used multivitamin
supplements in 1 year
before
diagnosis/interview,
area of study, and log-
converted amounts for
grains, fruit, vegetables,
total plants, total
carotenoids, folic acid,
dietary fibre, conjugated
linoleic acid, vitamin E,
vitamin C, retinol, total
fat, and linoleic acid
µg/day
<17.9
17.9-49.4
49.5-100
>100
Jian58
age at interview, BMI,
locality of residence,
µg/day
<70.7
70.7-126.8
126.9-
230.3
>230.3
Page | 48
education, family
income, marital status,
number of children,
family history of
prostate cancer, tea
drinking, total caloric
intake (kcal/day) and
total fat intake (gm/day
Kirsh65
age, total energy, race,
study centre, family
history of prostate
cancer, body mass
index, smoking status,
physical activity, total fat
intake, red meat intake,
history of diabetes,
aspirin use, and number
of screening
examinations during the
follow-up period
median µg/day
65
122
178
241
359
Lu51
age, race, education,
alcohol drinking, pack-
years of smoking, family
history of prostate
cancer, and total dietary
caloric intake
µg
<23.0516
23.0517-
71.1566
71.1567-
120.847
>120.848
McCann55
age, education, body
mass index, cigarette
smoking status, total
energy, veg intake
µg
≤36
36-99
99-201
>201
Rohrmann48
age, race or ethnicity,
cigarette smoking, BMI,
leisure-time physical
activity, alcohol
consumption, energy
intake, intake of protein,
and intake of
polyunsaturated fatty
acids
µg
11
33
56
93
171
Schuurman67
age, family history of
prostate cancer,
socioeconomic status,
and alcohol from white
or fortified wine
mg/day
0.012
0.045
0.1
0.2
0.4
Stram66
age, BMI, education and
family history of
prostate cancer
µg/1000kcal
≤19.5
19.6-48.1
48.2-91.7
91.8-189.9
≥190
Page | 49
4.3. Lutein
Reference
Adjustments
Measurement
Q1
Q2
Q3
Q4
Q5
Deneo-
Pellegrini60
age, residence,
urban/rural status,
education, family history
of prostate cancer, body
mass index and total
energy intake
µg/day
≤1214
1215–2086
2087–3593
3594+
Giovannucci68
energy-adjusted nutrient,
adjusted for age by
stratified analysis
µg
<1799
1799-2665
2666-3620
3621-5100
>5100
Jain61
log total energy,
vasectomy, age, ever-
smoked, marital status,
study area, body mass
index, education, ever-
used multivitamin
supplements in 1 yr.
before
diagnosis/interview, area
of study, and log-
converted amounts for
grains, fruit, vegetables,
total plants, total
carotenoids, folic acid,
dietary fibre, conjugated
linoleic acid, vitamin E,
vitamin C, retinol, total
fat, and linoleic acid
µg/day
<1019
1018-1653
1654-2684
>2684
Lu51
age, race, education,
alcohol drinking, pack-
years of smoking, family
history of prostate cancer,
and total dietary caloric
intake
µg
<1009.78
1009.79-
1666.75
1666.76-
2916.75
>2916.76
McCann55
age, education, body mass
index, cigarette smoking
status, total energy, veg
intake
µg/day
3029
3029-4975
4975-7168
>7168
Meyer62
age, education, family
history of prostate cancer,
group, dietary energy
not reported
Stram66
age, BMI, education and
family history of prostate
cancer
µg/1000kcal
≤594.4
594.5-
852.7
852.7-
1158.2
1158.3-
1661.3
≥1661.4
Page | 50
4.4. Lutein and Zeaxanthin
Reference
Adjustments
Measurement
Q1
Q2
Q3
Q4
Q5
Agalliu64
age at baseline, race, BMI,
exercise activity, and education
µg, median
values
1617.5
2220.1
2763.4
3506.2
5346.0
Bosetti56
Estimates from multiple logistic
regression models including terms
for age, study centre, education,
physical activity, body mass index,
family history of prostate cancer
and total calorie intake
µg, mean
values (SD)
Cohen59
fat, energy, race, age, family
history of prostate cancer, body
mass index, prostate-specific
antigen tests in previous 5 years,
and education
µg
<800
800-1299
1300-1999
≥2000
Hodge57
state, age group, year, country of
birth, socioeconomic group, and
family history of prostate cancer
µg/day
1101-
1531
1532-
1891
1892-2456
2457+
Jian58
age at interview, BMI, locality of
residence, education, family
income, marital status, number of
children, family history of
prostate cancer, tea drinking,
total caloric intake (kcal/day) and
total fat intake (gm/day
µg/day
<746.2
746.2-
1718.4
1718.5-
3590.6
>3590.6
Kirsh65
age, total energy, race, study
centre, family history of prostate
cancer, body mass index, smoking
status, physical activity, total fat
intake, red meat intake, history of
diabetes, aspirin use, and number
of screening examinations during
the follow-up period
median µg/day
1437
1995
2501
3138
4428
Rohrmann48
age, race or ethnicity, cigarette
smoking, BMI, leisure-time
physical activity, alcohol
consumption, energy intake,
intake of protein, and intake of
polyunsaturated fatty acids
µg/day
1308
2271
3184
4347
6788
Schuurman67
age, family history of prostate
cancer, socioeconomic status, and
alcohol from white or fortified
wine
mg/day
1.4
1.9
2.4
2.9
3.9
Page | 51
Appendix 5 – Contents of Do Files Used to Complete Meta-
analysis in Stata
Figure B - Do File Used for Case-control/NCC study analysis
Figure C - Do File Used for Cohort/Case-cohort Study Analysis
Page | 52
Appendix 6 – Full Results of Meta-Analyses of Dietary Intake
of Carotenoids and Prostate Cancer Incidence
6.1. α-Carotene
Table A - Full Results of Included Case-Control/NCC Studies
Reference
OR
95% CI
Log OR
SE
Weight (%)
Bosetti56
0.85
0.66-1.11
-0.1625189
0.1361627
9.96
Cohen59
0.75
0.51–1.09
-0.2876821
0.1907448
5.07
Deneo-Pellegrini60
0.9
0.5-1.6
-0.1053605
0.2935531
2.14
Hodge57
0.8
0.6-1.1
-0.2231435
0.1624764
6.99
Jain61
1.06
0.79-1.43
0.0582689
0.1527579
7.91
Jian58
0.43
0.21-0.85
-0.8439701
0.3476792
1.53
Lu51
0.47
0.14–1.66
-0.7550226
0.643796
0.45
McCann55
0.91
0.59-1.39
-0.0943106
0.2161298
3.95
Meyer62
1.00
0.53-1.89
0
0.3247841
1.75
Rohrmann48
0.96
0.87-1.07
-0.040822
0.0553473
60.25
Overall
0.915
0.841-0.996
-0.088831
-
100
Heterogeneity
Chi2 (d.f.)
p
I2 (%)
9.61 (9)
0.383
6.3
Test of Overall Effect
z
p
2.06
0.039
Figure D - Forrest Plot Showing Risk Estimates from Included Case–Control/NCC
Studies of Dietary α-Carotene Intake and Prostate Cancer Risk
Page | 53
Figure E - Funnel Plot Examining Publication Bias in Included Case-Control/NCC Studies
of Dietary α-Carotene and Prostate Cancer Risk
Table B - Full Results of Included Cohort/Case-Cohort Studies
Reference
HR
95% CI
log HR
SE
Weight (%)
Giovannucci68
1.09
0.87-1.36
0.0861777
0.1129117
12.56
Kirsh65
0.92
0.76-1.10
-0.0833816
0.0911693
19.26
Schuurman67
0.85
0.62-1.17
-0.1625189
0.1630217
6.02
Stram66
0.94
0.85-1.04
-0.0618754
0.0515796
60.17
Umesawa63
0.74
0.42–1.29
-0.3011051
0.2835445
1.99
Overall
0.943
0.872-1.020
-0.058689
-
100
Heterogeneity
Chi2 (d.f.)
p
I2 (%)
2.86 (4)
0.582
0.0
Test of Overall Effect
z
p
1.46
0.145
Page | 54
Figure F - Forrest Plot Showing Risk Estimates from Included Cohort/Case-Cohort
Studies of Dietary α-Carotene Intake and Prostate Cancer Risk
Figure G - Funnel Plot Examining Publication Bias in Included Cohort/Case-Cohort
Studies of Dietary α-Carotene and Prostate Cancer Risk
Page | 55
6.2. β-Cryptoxanthin
Table C - Full Results of Included Case-Control/NCC Studies
Reference
OR
95% CI
Log OR
SE
Weight (%)
Bosetti56
0.9
0.69–1.16
-0.1053605
0.1294799
11.31
Cohen59
0.93
0.64-1.36
-0.0725707
0.1939058
5.04
Hodge57
0.9
0.7–1.3
-0.1053605
0.1876147
5.38
Jain61
1.44
1.09-1.89
0.3646432
0.1387417
9.85
Jian58
0.15
0.06-0.34
-1.89712
0.4175053
1.09
Lu51
0.92
0.26-3.2
-0.0833816
0.6359859
0.47
McCann55
0.92
0.64-1.33
-0.0833816
0.1880411
5.36
Rohrmann48
0.87
0.79-0.97
-0.1392621
0.0555117
61.51
Overall
0.908
0.834-0.989
-0.096511
-
100
Heterogeneity
Chi2 (d.f.)
p
I2 (%)
30.27
0.000
76.9
Test of Overall Effect
z
p
2.22
0.027
Figure H - Forrest Plot Showing Risk Estimates from Included Case–Control/NCC
Studies of Dietary β-Cryptoxanthin Intake and Prostate Cancer Risk
Page | 56
Figure I - Funnel Plot Examining Publication Bias in Included Case-Control/NCC Studies
of Dietary β-Cryptoxanthin and Prostate Cancer Risk
Table D - Full Results of Included Cohort/Case-Cohort Studies
Reference
HR
95% CI
Log HR
SE
Weight (%)
Agalliu64
0.94
0.69–1.28
-.0618754
0.1575181
6.27
Giovannucci68
0.94
0.75-1.17
-.0618754
0.111673
12.47
Kirsh65
1.05
0.87-1.27
.0487901
0.0970545
16.51
Schuurman67
1.41
1.03-1.92
0.3435897
0.1575181
6.27
Stram66
0.94
0.85-1.04
-.0618754
0.0515796
58.47
Overall
0.982
0.909-1.061
-0.018163
-
100
Heterogeneity
Chi2
(d.f.)
p
I2 (%)
6.70
0.153
40.3
Test of Overall Effect
z
p
0.46
0.645
Page | 57
Figure J - Forrest Plot Showing Risk Estimates from Included Cohort/Case-Cohort
Studies of Dietary β-Cryptoxanthin Intake and Prostate Cancer Risk
Figure K - Funnel Plot Examining Publication Bias in Included Cohort/Case-Cohort
Studies of Dietary β-Cryptoxanthin and Prostate Cancer Risk
Page | 58
6.3. Lutein
Table E - Full Results of Included Case-Control/NCC Studies
Reference
OR
95% CI
log OR
SE
Weight (%)
Deneo-Pellegrini60
0.7
0.4-1.3
-0.356675
0.3158363
15.68
Jain61
0.81
0.65-1.18
-0.210721
0.1919568
42.46
Lu51
0.55
0.16-1.88
-0.597837
0.6270963
3.98
McCann55
0.71
0.43-1.16
-0.3424903
0.2504644
24.94
Meyer62
0.86
0.44-1.7
-0.1508229
0.3476792
12.94
Overall
0.760
0.595-0.971
-2.744367
-
100
Heterogeneity
Chi2
(d.f.)
p
I2 (%)
0.64 (4)
0.958
0.0
Test of Overall Effect
z
p
2.19
0.028
Figure L - Forrest Plot Showing Risk Estimates from Included Case–Control/NCC
Studies of Dietary Lutein Intake and Prostate Cancer Risk
Page | 59
Figure M - Funnel Plot Examining Publication Bias in Included Case-Control/NCC
Studies of Lutein and Prostate Cancer Risk
Table F - Full Results of Included Cohort/Case-Cohort Studies
Reference
HR
95% CI
log HR
SE
Weight (%)
Giovannucci68
1.1
0.88-1.37
0.0953102
0.1119901
19.02
Stram66
0.98
0.88–1.09
-0.0202027
0.0542757
80.98
Overall
1.002
0.910-1.102
0.001998
-
100
Heterogeneity
Chi2
(d.f.)
P
I2 (%)
0.86 (1)
0.353
0.0
Test of Overall Effect
z
P
0.04
0.971
Page | 60
Figure M - Forrest Plot Showing Risk Estimates from Included Cohort/Case-Cohort
Studies of Dietary Lutein Intake and Prostate Cancer Risk
Figure N - Funnel Plot Examining Publication Bias in Included Cohort/Case-Cohort
Studies of Dietary Lutein and Prostate Cancer Risk
Page | 61
6.4. Lutein & Zeaxanthin
Table G - Full Results of Included Case-Control/NCC Studies
Reference
OR
95% CI
log OR
SE
Weight (%)
Bosetti56
0.91
0.69-1.2
-0.0943106
0.1411389
9.14
Cohen59
0.68
0.45-1
-0.3856625
0.1967666
4.70
Hodge57
0.9
0.7-1.3
-0.1053605
0.1876147
5.17
Jian58
0.02
0.01-0.1
-3.912023
0.8211418
0.27
Rohrmann48
0.82
0.74-0.9
-0.198451
0.0474951
80.71
Overall
0.816
0.751-0.888
-0.203341
-
100
Heterogeneity
Chi2 (d.f.)
p
I2 (%)
22.14 (4)
0.000
81.9
Test of Overall Effect
Z
p
4.76
0.000
Figure O - Forrest Plot Showing Risk Estimates from Included Case–Control/NCC
Studies of Dietary Lutein & Zeaxanthin Intake and Prostate Cancer Risk
Page | 62
Figure P - Funnel Plot Examining Publication Bias in Included Case-Control/NCC Studies
of Dietary Lutein & Zeaxanthin and Prostate Cancer Risk
Table H - Full Results of Included Cohort/Case-Cohort Studies
Reference
HR
95% CI
log HR
SE
Weight (%)
Agalliu64
0.97
0.72-1.3
-0.0304592
0.1493997
22.35
Kirsh65
0.95
0.78-1.14
-0.0512933
0.0930212
57.64
Schuurman67
0.91
0.66-1.24
-0.0943106
0.1578684
20.01
Overall
0.946
0.824-1.087
-0.055513
-
100
Heterogeneity
Chi2 (d.f.)
P
I2 (%)
0.09 (2)
0.956
0.0
Test of Overall Effect
z
P
0.78
0.434
Page | 63
Figure Q - Forrest Plot Showing Risk Estimates from Included Cohort/Case-Cohort
Studies of Dietary Lutein & Zeaxanthin Intake and Prostate Cancer Risk
Figure R - Funnel Plot Examining Publication Bias in Included Cohort/Case-Cohort
Studies of Dietary Lutein & Zeaxanthin and Prostate Cancer Risk
Page | 64
Appendix 7 – Meta-Analysis of Case-Control/Nested Case-
Control Studies of Dietary Carotenoids and Prostate Cancer
Risk Excluding Results from Jian et al
7.1. α-Carotene
Table I - Full Results of Included Studies
Reference
OR
95% CI
Log OR
SE
Weight (%)
Bosetti56
0.85
0.66-1.11
-0.1625189
0.1361627
10.11
Cohen59
0.75
0.51–1.09
-0.2876821
0.1907448
5.15
Deneo-Pellegrini60
0.9
0.5-1.6
-0.1053605
0.2935531
2.18
Hodge57
0.8
0.6-1.1
-0.2231435
0.1624764
7.10
Jain61
1.06
0.79-1.43
0.0582689
0.1527579
8.03
Lu51
0.47
0.14–1.66
-0.7550226
0.643796
0.45
McCann55
0.91
0.59-1.39
-0.0943106
0.2161298
4.01
Meyer62
1
0.53-1.89
0
0.3247841
1.78
Rohrmann48
0.96
0.87-1.07
-0.040822
0.0553473
61.19
Overall
0.926
0.851-1.008
-0.076881
-
100
Heterogeneity
Chi2 (d.f.)
P
I2 (%)
4.82 (8)
0.777
0.0
Test of Overall
Effect
Z
P
1.78
0.076
Figure S - Forrest Plot Showing Risk Estimates from Included Studies of Dietary α-
Carotene Intake and Prostate Cancer Risk
Page | 65
Figure T - Funnel Plot Examining Publication Bias in Included Studies of Dietary α-
Carotene and Prostate Cancer Risk
7.2. β-Cryptoxanthin
Table J - Full Results of Included Studies
Reference
OR
95% CI
log OR
SE
Weight (%)
Bosetti56
0.9
0.69–1.16
-0.1053605
0.1294799
11.43
Cohen59
0.93
0.64-1.36
-0.0725707
0.1939058
5.10
Hodge57
0.9
0.7–1.3
-0.1053605
0.1876147
5.44
Jain61
1.44
1.09-1.89
0.3646432
0.1387417
9.95
Lu51
0.92
0.26-3.2
-0.0833816
0.6359859
0.47
McCann55
0.92
0.64-1.33
-0.0833816
0.1880411
5.42
Rohrmann48
0.87
0.79-0.97
-0.1392621
0.0555117
62.18
Overall
0.926
0.850-1.009
-0.076881
-
100
Heterogeneity
Chi2 (d.f.)
P
I2 (%)
11.46 (6)
0.075
47.7
Test of Overall
Effect
z
P
1.75
0.080
Page | 66
Figure U - Forrest Plot Showing Risk Estimates from Included Studies of Dietary β-
Cryptoxanthin Intake and Prostate Cancer Risk
Figure V - Funnel Plot Examining Publication Bias in Included Studies of Dietary β-
Cryptoxanthin and Prostate Cancer Risk
Page | 67
7.3. Lutein & Zeaxanthin
Table K - Full Results of Included Studies
Reference
OR
95% CI
log OR
SE
Weight (%)
Bosetti56
0.91
0.69-1.2
-0.0943106
0.1411389
9.16
Cohen59
0.68
0.45-1
-0.3856625
0.1967666
4.72
Hodge57
0.9
0.7-1.3
-0.1053605
0.1876147
5.19
Rohrmann48
0.82
0.74-0.9
-0.198451
0.0474951
80.93
Overall
0.825
0.758-0.897
-0.1923729
-
100
Heterogeneity
Chi2 (d.f.)
P
I2 (%)
1.68 (3)
0.642
0.0
Test of Overall Effect
z
P
4.51
0.000
Figure W - Forrest Plot Showing Risk Estimates from Included Studies of Dietary lutein
& Zeaxanthin Intake and Prostate Cancer Risk
Page | 68
Figure X - Funnel Plot Examining Publication Bias in Included Studies of Dietary lutein
& Zeaxanthin and Prostate Cancer Risk
Page | 69
References
1. Ferlay J, Soerjomataram I, Ervik M, Dikshit R, Eser S, Mathers C, Rebelo M, Parkin DM, Forman D, Bray, F. GLOBOCAN 2012 v1.0,
Cancer Incidence and Mortality Worldwide: IARC CancerBase No. 11 [Internet]. Lyon, France: International Agency for Research
on Cancer; 2013. Available from: http://globocan.iarc.fr, accessed on 12/03/2014.
2. Ferlay J, Steliarova-Foucher E, Lortet-Tieulent J, Rosso S, Coebergh JWW, Comber H, Forman D, Bray F. Cancer incidence and
mortality patterns in Europe: estimates for 40 countries in 2012. Eur J Cancer. 2013 Apr; 49(6):1374-403.
3. Etzioni, R., Penson, D. F., Legler, J. M., di Tommaso, D., Boer, R., Gann, P. H., & Feuer, E. J. (2002). Overdiagnosis due to prostate-
specific antigen screening: lessons from US prostate cancer incidence trends. Journal of the National Cancer Institute, 94(13), 981-
990.
4. Grönberg, H. (2003). Prostate cancer epidemiology. The Lancet, 361(9360), 859-864.
5. Rowley, K. H.M., & Mason, M.D. (1997). The aetiology and pathogenesis of prostate cancer. Clinical Oncology, 9(4), 213-218.
6. Schrecengost, R., & Knudsen, K. E. (2013, June). Molecular pathogenesis and progression of prostate cancer. In Seminars in
oncology (Vol. 40, No. 3, pp. 244-258).
7. Perry, A. S., Watson, R. W. G., Lawler, M., & Hollywood, D. (2010). The epigenome as a therapeutic target in prostate
cancer. Nature Reviews Urology, 7(12), 668-680.
8. Ørsted, D. D., & Bojesen, S. E. (2013). The link between benign prostatic hyperplasia and prostate cancer. Nature Reviews
Urology, 10(1), 49-54.
9. Damber, J-E. Gunnar, A. (2008, May). Prostate cancer. The Lancet (Vol. 371, Issue 9625) 1710-1721.
10. Kamangar, F., Dores, G. M., & Anderson, W. F. (2006). Patterns of cancer incidence, mortality, and prevalence across five
continents: defining priorities to reduce cancer disparities in different geographic regions of the world. Journal of clinical
oncology, 24(14), 2137-2150.
11. Minami, Y., Staples, M. P., & Giles, G. G. (1993). The incidence of colon, breast and prostate cancer in Italian migrants to Victoria,
Australia. European Journal of Cancer, 29(12), 1735-1740.
12. Parkin, D.M., Boyd, L., & Walker, L.C. (2011). The fraction of cancer attributable to lifestyle and environmental factors in the UK in
2010. British Journal of Cancer, 105, S77–S81.
13. Zu, K., & Giovannucci, E. (2009). Smoking and aggressive prostate cancer: a review of the epidemiologic evidence. Cancer causes
& control, 20(10), 1799-1810.
14. Allott, E. H., Masko, E. M., & Freedland, S. J. (2013). Obesity and prostate cancer: weighing the evidence. European urology, 63(5),
800-809.
15. Bonn, S.E., Sjölander, A., Lagerros, Y.T., Wiklund, F., Stattin, P., Holmberg, E., Grönberg, H., Bälter, K. (2014). Physical Activity and
Survival among Men Diagnosed with Prostate Cancer. [Unpublished].
16. Key, T. J., Allen, N., Appleby, P., Overvad, K., Tjønneland, A., Miller, A., ... & Riboli, E. (2004). Fruits and vegetables and prostate
cancer: No association among 1,104 cases in a prospective study of 130,544 men in the European Prospective Investigation into
Cancer and Nutrition (EPIC). International journal of cancer, 109(1), 119-124.
17. Hwang, Y. W., Kim, S. Y., Jee, S. H., Kim, Y. N., & Nam, C. M. (2009). Soy food consumption and risk of prostate cancer: a meta-
analysis of observational studies. Nutrition and cancer, 61(5), 598-606.
18. Venkateswaran, V., & Klotz, L. H. (2010). Diet and prostate cancer: mechanisms of action and implications for
chemoprevention. Nature Reviews Urology, 7(8), 442-453.
19. Liu, B., Mao, Q., Cao, M., & Xie, L. (2012). Cruciferous vegetables intake and risk of prostate cancer: A meta‐analysis. International
Journal of Urology, 19(2), 134-141.
20. Szymanski, K. M., Wheeler, D. C., & Mucci, L. A. (2010). Fish consumption and prostate cancer risk: a review and meta-
analysis. The American journal of clinical nutrition, 92(5), 1223-1233.
21. Cao, S., Liu, L., Yin, X., Wang, Y., Liu, J., & Lu, Z. (2014). Coffee consumption and risk of prostate cancer: a meta-analysis of
prospective cohort studies. Carcinogenesis, 35(2), 256-261.
22. Huncharek, M., Muscat, J., & Kupelnick, B. (2008). Dairy products, dietary calcium and vitamin D intake as risk factors for prostate
cancer: a meta-analysis of 26,769 cases from 45 observational studies. Nutrition and cancer, 60(4), 421-441.
23. Dennis, L. K., Snetselaar, L. G., Smith, B. J., Stewart, R. E., & Robbins, M. E. (2004). Problems with the assessment of dietary fat in
prostate cancer studies. American journal of epidemiology, 160(5), 436-444.
24. Xu, X., Cheng, Y., Li, S., Zhu, Y., Xu, X., Zheng, X., ... & Xie, L. (2014). Dietary carrot consumption and the risk of prostate
cancer. European journal of nutrition, 1-9.
25. Tanumihardjo, S. A. (2012). Provitamin A Carotenoids and Cancer Prevention. Carotenoids and Human Health (pp. 182). Retrieved
from https://www.google.ie/search?tbm=bks&hl=en&q=Carotenoids+and+Human+Health.
26. Krinsky, N. I., & Johnson, E. J. (2005). Carotenoid actions and their relation to health and disease. Molecular aspects of
medicine, 26(6), 459-516.
27. Sommer, A., & Vyas, K. S. (2012). A global clinical view on vitamin A and carotenoids. The American journal of clinical
nutrition, 96(5), 1204S-1206S.
28. National Academy of Sciences, Institute of Medicine, Food and Nutrition Board. (2000). Dietary Reference Intakes for Vitamin C,
Vitamin E, Selenium, and Carotenoids (pp. 326). Retrieved from http://fnic.nal.usda.gov/dietary-guidance/dri-reports/vitamin-c-
vitamin-e-selenium-and-carotenoids
29. Britton, G., Liaaen-Jensen, S., & Pfander, H. (2009). Supplements. Carotenoids Volume 5: Nutrition and Health (pp. 79). Retrieved
from http://books.google.ie/books?id=c8USiLi73dUC&printsec=frontcover#v=onepage&q&f=false.
Page | 70
30. Hardin, J., Cheng, I., & Witte, J. S. (2011). Impact of consumption of vegetable, fruit, grain, and high glycemic index foods on
aggressive prostate cancer risk. Nutrition and cancer, 63(6), 860-872.
31. Wertz, K. (2009). Lycopene effects contributing to prostate health. Nutrition and cancer, 61(6), 775-783.
32. Chen, J., Song, Y., & Zhang, L. (2012). Lycopene/Tomato consumption and the risk of prostate cancer: a systematic review and
meta-analysis of prospective studies. Journal of nutritional science and vitaminology, 59(3), 213-223.
33. Albanes, D., Heinonen, O. P., Huttunen, J. K., Taylor, P. R., Virtamo, J., Edwards, B. K., ... & Palmgren, J. (1995). Effects of alpha-
tocopherol and beta-carotene supplements on cancer incidence in the Alpha-Tocopherol Beta-Carotene Cancer Prevention
Study. The American journal of clinical nutrition, 62(6), 1427S-1430S.
34. World Cancer Research Fund / American Institute for Cancer Research. Food, Nutrition, Physical Activity, and the Prevention of
Cancer: a Global Perspective. Washington DC: AICR, 2007
35. National Nutrient Database for Standard Reference, Release 26, Software v.1.3.1. Retrieved at
http://ndb.nal.usda.gov/ndb/nutrients/index
36. Li, C., Ford, E. S., Zhao, G., Balluz, L. S., Giles, W. H., & Liu, S. (2011). Serum α-Carotene Concentrations and Risk of Death Among
US Adults: The Third National Health and Nutrition Examination Survey Follow-up Study. Archives of internal medicine, 171(6),
507-515.
37. Levy, J., Bosin, E., Feldman, B., Giat, Y., Miinster, A., Danielko, M., Sharoni, Y. (1995). Lycopene is a more potent inhibitor of
human cancer cell proliferation than either α‐carotene or β‐carotene. Nutrition and Cancer, Vol. 24, Iss. 3.
38. Lorenzo, Y., Azquetal, A., Luna, L., Bonilla, F., Domínguez, G., Collins, A. (2009). The carotenoid β-cryptoxanthin stimulates the
repair of DNA oxidation damage in addition to acting as an antioxidant in human cells. Carcinogenesis, 30 (2):308-314.
39. Chang, S., Erdman, Jr, J. W., Clinton, S. K., Vadiveloo, M., Strom, S. S., Yamamura, Y., ... & Hursting, S. D. (2005). Relationship
between plasma carotenoids and prostate cancer. Nutrition and cancer, 53(2), 127-134.
40. Zhang, J., Dhakal, I., Stone, A., Ning, B., Greene, G., Lang, N. P., & Kadlubar, F. F. (2007). Plasma carotenoids and prostate cancer: a
population-based case-control study in Arkansas. Nutrition and cancer, 59(1), 46-53.
41. Nomura, A. M., Stemmermann, G. N., Lee, J., & Craft, N. E. (1997). Serum micronutrients and prostate cancer in Japanese
Americans in Hawaii. Cancer Epidemiology Biomarkers & Prevention, 6(7), 487-491.
42. Huang, H. Y., Alberg, A. J., Norkus, E. P., Hoffman, S. C., Comstock, G. W., & Helzlsouer, K. J. (2003). Prospective study of
antioxidant micronutrients in the blood and the risk of developing prostate cancer. American journal of epidemiology, 157(4),
335-344.
43. Gunasekera, R. S., Sewgobind, K., Desai, S., Dunn, L., Black, H. S., McKeehan, W. L., & Patil, B. (2007). Lycopene and lutein inhibit
proliferation in rat prostate carcinoma cells. HNUC, 58(2), 171-177.
44. Age-Related Eye Disease Study Research Group. (2007). The relationship of dietary carotenoid and vitamin A, E, and C intake with
age-related macular degeneration in a case-control study: AREDS Report No. 22. Archives of ophthalmology, 125(9), 1225.
45. Richer, S., Stiles, W., Statkute, L., Pulido, J., Frankowski, J., Rudy, D., ... & Nyland, J. (2004). Double-masked, placebo-controlled,
randomized trial of lutein and antioxidant supplementation in the intervention of atrophic age-related macular degeneration: the
Veterans LAST study (Lutein Antioxidant Supplementation Trial). Optometry-Journal of the American Optometric
Association, 75(4), 216-229.
46. Seddon, J. M., Ajani, U. A., Sperduto, R. D., Hiller, R., Blair, N., Burton, T. C., ... & Willett, W. (1994). Dietary carotenoids, vitamins
A, C, and E, and advanced age-related macular degeneration. Jama, 272(18), 1413-1420.
47. European Commission for Food and Feed Safety. (n.d.). Database on Food Additives. Retrieved March 13, 2014, from
https://webgate.ec.europa.eu/sanco_foods/main/index.cfm - See more at: http://reffor.us/index.php#sthash.KvV2eGaM.dpuf
48. Rohrmann, S., Giovannucci, E., Willett, W. C., & Platz, E. A. (2007). Fruit and vegetable consumption, intake of micronutrients, and
benign prostatic hyperplasia in US men. The American journal of clinical nutrition, 85(2), 523-529.
49. Abdel-Aal, E. S. M., Akhtar, H., Zaheer, K., & Ali, R. (2013). Dietary sources of lutein and zeaxanthin carotenoids and their role in
eye health. Nutrients, 5(4), 1169-1185.
50. Maccarrone, M., Bari, M., Gasperi, V., & Demmig-Adams, B. (2005). The photoreceptor protector zeaxanthin induces cell death in
neuroblastoma cells. Anticancer research, 25(6B), 3871-3876.
51. Lu, Q. Y., Hung, J. C., Heber, D., Go, V. L. W., Reuter, V. E., Cordon-Cardo, C., ... & Zhang, Z. F. (2001). Inverse associations between
plasma lycopene and other carotenoids and prostate cancer. Cancer Epidemiology Biomarkers & Prevention, 10(7), 749-756
52. Holden, J. M., Eldridge, A. L., Beecher, G. R., Marilyn Buzzard, I., Bhagwat, S., Davis, C. S., ... & Schakel, S. (1999). Carotenoid
content of US foods: an update of the database. Journal of Food Composition and Analysis, 12(3), 169-196.
53. Key, T. J., Appleby, P. N., Allen, N. E., Travis, R. C., Roddam, A. W., Jenab, M., ... & Riboli, E. (2007). Plasma carotenoids, retinol,
and tocopherols and the risk of prostate cancer in the European Prospective Investigation into Cancer and Nutrition study. The
American journal of clinical nutrition, 86(3), 672-681.
54. Moher, D., Liberati, A., Tetzlaff, J., & Altman, D. G. (2009). Preferred reporting items for systematic reviews and meta-analyses:
the PRISMA statement. Annals of internal medicine, 151(4), 264-269.
55. McCann, S. E., Ambrosone, C. B., Moysich, K. B., Brasure, J., Marshall, J. R., Freudenheim, J. L., ... & Graham, S. (2005). Intakes of
selected nutrients, foods, and phytochemicals and prostate cancer risk in western New York. Nutrition and cancer, 53(1), 33-41.
56. Bosetti, C., Talamini, R., Montella, M., Negri, E., Conti, E., Franceschi, S., & La Vecchia, C. (2004). Retinol, carotenoids and the risk
of prostate cancer: A case‐control study from Italy. International journal of cancer, 112(4), 689-692.
57. Hodge, A. M., English, D. R., McCredie, M. R., Severi, G., Boyle, P., Hopper, J. L., & Giles, G. G. (2004). Foods, nutrients and
prostate cancer. Cancer Causes & Control, 15(1), 11-20.
Page | 71
58. Jian, L., Du, C. J., Lee, A. H., & Binns, C. W. (2005). Do dietary lycopene and other carotenoids protect against prostate
cancer? International Journal of Cancer, 113(6), 1010-1014.
59. Cohen, J. H., Kristal, A. R., & Stanford, J. L. (2000). Fruit and vegetable intakes and prostate cancer risk. Journal of the National
Cancer Institute, 92(1), 61-68.
60. Deneo-Pellegrini, H., De Stefani, E., Ronco, A., & Mendilaharsu, M. (1999). Foods, nutrients and prostate cancer: a case–control
study in Uruguay. British journal of cancer, 80(3-4), 591.
61. Jain, M. G., Hislop, G. T., Howe, G. R., & Ghadirian, P. (1999). Plant foods, antioxidants, and prostate cancer risk: findings from
case-control studies in Canada. Nutrition and cancer, 34(2), 173-184.
62. Meyer, F., Bairati, I., Fradet, Y., & Moore, L. (1997). Dietary energy and nutrients in relation to preclinical prostate cancer.
63. Umesawa, M., Iso, H., Mikami, K., Kubo, T., Suzuki, K., Watanabe, Y., ... & Tamakoshi, A. (2013). Relationship between vegetable
and carotene intake and risk of prostate cancer: the JACC study. British journal of cancer, 110, 792-796.
64. Agalliu, I., Kirsh, V. A., Kreiger, N., Soskolne, C. L., & Rohan, T. E. (2011). Oxidative balance score and risk of prostate cancer:
Results from a case-cohort study. Cancer epidemiology, 35(4), 353-361.
65. Kirsh, V., Hayes, R., Mayne, S., Chatterjje, N., Subar, A., Dixon, L.B., Albanes, D., Andriole, G., Urban, D., Peters, U. (2006).
Supplemental and Dietary Vitamin E, β-Carotene, and Vitamin C Intakes and Prostate Cancer Risk. JNCI J Natl Cancer
Inst, 98 (4): 245-254.
66. Stram, D. O., Hankin, J. H., Wilkens, L. R., Park, S., Henderson, B. E., Nomura, A. M., ... & Kolonel, L. N. (2006). Prostate cancer
incidence and intake of fruits, vegetables and related micronutrients: the multiethnic cohort study*(United States). Cancer Causes
& Control, 17(9), 1193-1207.
67. Schuurman, A. G., Goldbohm, R. A., Brants, H. A., & van den Brandt, P. A. (2002). A prospective cohort study on intake of retinol,
vitamins C and E, and carotenoids and prostate cancer risk (Netherlands). Cancer Causes & Control, 13(6), 573-582.
68. Giovannucci, E., Ascherio, A., Rimm, E. B., Stampfer, M. J., Colditz, G. A., & Willett, W. C. (1995). Intake of carotenoids and retinol
in relation to risk of prostate cancer. Journal of the national cancer institute, 87(23), 1767-1776.
69. Langholz, B. (2005). Case–Control Study, Nested. In Peter Armitage & Theodore Colton (Ed.), Encyclopedia of Biostatistics (2nd
ed., pp. 646–655). Retrieved from http://hydra.usc.edu/pm518b/literature/eob_nested.pdf.
70. Kang, S., & Cai, J. (2009). Marginal hazards model for case-cohort studies with multiple disease outcomes. Biometrika, 96(4), 887-
901.
71. Lin, Y. W., Hu, Z. H., Wang, X., Mao, Q. Q., Qin, J., Zheng, X. Y., & Xie, L. P. (2014). Tea consumption and prostate cancer: an
updated meta-analysis. World journal of surgical oncology, 12(1), 38.
72. Westreich, D. (2012). Berkson’s bias, selection bias, and missing data. Epidemiology (Cambridge, Mass.), 23(1), 159.
73. Petrisor, B. A., & Bhandari, M. (2007). The hierarchy of evidence: levels and grades of recommendation. Indian journal of
orthopaedics, 41(1), 11.
74. Kolonel, L. N., Henderson, B. E., Hankin, J. H., Nomura, A. M., Wilkens, L. R., Pike, M. C., ... & Nagamine, F. S. (2000). A multiethnic
cohort in Hawaii and Los Angeles: baseline characteristics. American journal of epidemiology, 151(4), 346-357.
75. Coughlin, S. S. (1990). Recall bias in epidemiologic studies. Journal of clinical epidemiology, 43(1), 87-91.
76. Patterson, R. E., Neuhouser, M. L., Hedderson, M. M., Schwartz, S. M., Standish, L. J., & Bowen, D. J. (2003). Changes in diet,
physical activity, and supplement use among adults diagnosed with cancer. Journal of the American Dietetic Association, 103(3),
323-328.
77. Mangels, A. R., Holden, J. M., Beecher, G. R., Forman, M. R., & Lanza, E. (1993). Carotenoid content of fruits and vegetables: an
evaluation of analytic data. Journal of the American Dietetic Association, 93(3), 284-296.
78. Nutrition Coordinating Center: Nutrition Data System. Minneapolis (MN): University of Minnesota. Accessed at
http://www.ncc.umn.edu/products/ndsr.html
79. Leitzmann, M. F., & Rohrmann, S. (2011). Risk factors for the onset of prostatic cancer: age, location, and behavioral
correlates. Clinical epidemiology, 4, 1-11.
80. Tamakoshi, A. (2007). Overview of the Japan Collaborative Cohort Study for evaluation of cancer (JACC). Asian Pac J Cancer Prev, 8
(Suppl), 1-8.
81. Kruep, E. J., Goodwin, B. B., & Chaudhari, S. (2013). Evaluation of Recent Trends in Treatment Patterns Among Men With Benign
Prostatic Hyperplasia. American journal of men's health, 7(3), 214-219.
82. Schenk, J. M., Kristal, A. R., Arnold, K. B., Tangen, C. M., Neuhouser, M. L., Lin, D. W., ... & Thompson, I. M. (2011). Association of
symptomatic benign prostatic hyperplasia and prostate cancer: results from the prostate cancer prevention trial. American
journal of epidemiology, 173(12), 1419-1428.
83. Smith, A. B., & Carson, C. C. (2009). Finasteride in the treatment of patients with benign prostatic hyperplasia: a
review. Therapeutics and clinical risk management, 5, 535.
84. Sarvis, J. A., & Thompson, I. M. (2008). Prostate cancer chemoprevention: update of the prostate cancer prevention trial findings
and implications for clinical practice. Current oncology reports, 10(6), 529-532.
85. Schwarz, S., Obermüller-Jevic, U. C., Hellmis, E., Koch, W., Jacobi, G., & Biesalski, H. K. (2008). Lycopene inhibits disease
progression in patients with benign prostate hyperplasia. The Journal of nutrition, 138(1), 49-53.
86. Müller, K. F., Briel, M., D’Amario, A., Kleijnen, J., Marusic, A., Wager, E., ... & Bassler, D. (2013). Defining publication bias: protocol
for a systematic review of highly cited articles and proposal for a new framework. Systematic reviews, 2(1), 34.
87. Begg, C. B., & Mazumdar, M. (1994). Operating characteristics of a rank correlation test for publication bias. Biometrics, 1088-
1101.
Page | 72
88. Egger, M., Smith, G. D., Schneider, M., & Minder, C. (1997). Bias in meta-analysis detected by a simple, graphical
test. Bmj, 315(7109), 629-634.
89. Stratton, J., & Godwin, M. (2011). The effect of supplemental vitamins and minerals on the development of prostate cancer: a
systematic review and meta-analysis. Family practice, 28(3), 243-252.
90. Posadzki, P., Lee, M. S., Onakpoya, I., Lee, H. W., Ko, B. S., & Ernst, E. (2013). Dietary supplements and prostate cancer: a
systematic review of double-blind, placebo-controlled randomised clinical trials. Maturitas, 75(2), 125-130.
91. Ilic, D., Forbes, K. M., & Hassed, C. (2011). Lycopene for the prevention of prostate cancer. Cochrane Database Syst Rev, 11.
92. Hwang, E. S., Stacewicz‐Sapuntzakis, M., & Bowen, P. E. (2012). Effects of heat treatment on the carotenoid and tocopherol
composition of tomato. Journal of food science, 77(10), C1109-C1114.
93. Burri, B. J., Burri, B. J., Chapman, M. H., Neidlinger, T. R., Seo, J. S., Ishida, B. K., ... & Ishida, B. K. (2008). Tangerine tomatoes
increase total and tetra-cis-lycopene isomer concentrations more than red tomatoes in healthy adult humans. International
journal of food sciences and nutrition, 60 (S1), 1-16.
94. Unlu, N. Z., Bohn, T., Clinton, S. K., & Schwartz, S. J. (2005). Carotenoid absorption from salad and salsa by humans is enhanced by
the addition of avocado or avocado oil. The Journal of nutrition, 135(3), 431-436.
95. McCullough, M. L., Feskanich, D., Stampfer, M. J., Giovannucci, E. L., Rimm, E. B., Hu, F. B., ... & Willett, W. C. (2002). Diet quality
and major chronic disease risk in men and women: moving toward improved dietary guidance. The American journal of clinical
nutrition, 76(6), 1261-1271
96. Lichtenstein, P., Holm, N. V., Verkasalo, P. K., Iliadou, A., Kaprio, J., Koskenvuo, M., ... & Hemminki, K. (2000). Environmental and
heritable factors in the causation of cancer—analyses of cohorts of twins from Sweden, Denmark, and Finland. New England
Journal of Medicine, 343(2), 78-85.
97. Jack, R. H., Davies, E. A., & Møller, H. (2010). Prostate cancer incidence, stage at diagnosis, treatment and survival in ethnic groups
in South‐East England. BJU international, 105(9), 1226-1230.
98. Michaud, D. S., Giovannucci, E. L., Ascherio, A., Rimm, E. B., Forman, M. R., Sampson, L., & Willett, W. C. (1998). Associations of
plasma carotenoid concentrations and dietary intake of specific carotenoids in samples of two prospective cohort studies using a
new carotenoid database. Cancer Epidemiology Biomarkers & Prevention, 7(4), 283-290.
99. Chung, H. Y., Ferreira, A. L. A., Epstein, S., Paiva, S. A., Castaneda-Sceppa, C., & Johnson, E. J. (2009). Site-specific concentrations
of carotenoids in adipose tissue: relations with dietary and serum carotenoid concentrations in healthy adults. The American
journal of clinical nutrition, 90(3), 533-539.
100. Scarmo, S., Cartmel, B., Lin, H., Leffell, D. J., Welch, E., Bhosale, P., ... & Mayne, S. T. (2010). Significant correlations of dermal total
carotenoids and dermal lycopene with their respective plasma levels in healthy adults. Archives of biochemistry and
biophysics, 504(1), 34-39.
101. Sen, A., Ren, J., Ruffin, M. T., Turgeon, D. K., Brenner, D. E., Sidahmed, E., ... & Djuric, Z. (2013). Relationships between Serum and
Colon Concentrations of Carotenoids and Fatty Acids in Randomized Dietary Intervention Trial. Cancer Prevention Research, 6(6),
558-565.
102. Clinton, S. K., Emenhiser, C., Schwartz, S. J., Bostwick, D. G., Williams, A. W., Moore, B. J., & Erdman, J. W. (1996). cis-trans
lycopene isomers, carotenoids, and retinol in the human prostate. Cancer Epidemiology Biomarkers & Prevention, 5(10), 823-833.
103. Dewell, A., Weidner, G., Sumner, M. D., Chi, C. S., & Ornish, D. (2008). A very-low-fat vegan diet increases intake of protective
dietary factors and decreases intake of pathogenic dietary factors. Journal of the American Dietetic Association, 108(2), 347-356.
104. McEvoy, C. T., Temple, N., & Woodside, J. V. (2012). Vegetarian diets, low-meat diets and health: a review. Public health
nutrition, 15(12), 2287-2294.
105. Alexander, D. D., Mink, P. J., Cushing, C. A., & Sceurman, B. (2010). A review and meta-analysis of prospective studies of red and
processed meat intake and prostate cancer. Nutrition journal, 9(1), 50.
106. Xie, B., & He, H. (2012). No association between egg intake and prostate cancer risk: a meta-analysis. Asian Pac J Cancer Prev, 13,
4677-4681.
107. Al-Delaimy, W. K., Slimani, N., Ferrari, P., Key, T., Spencer, E., Johansson, I., ... & Riboli, E. (2005). Plasma carotenoids as
biomarkers of intake of fruits and vegetables: ecological-level correlations in the European Prospective Investigation into Cancer
and Nutrition (EPIC). European journal of clinical nutrition, 59(12), 1397-1408.
108. Parsons, J. K., Newman, V. A., Mohler, J. L., Pierce, J. P., Flatt, S., & Marshall, J. (2008). Dietary modification in patients with
prostate cancer on active surveillance: a randomized, multicentre feasibility study. BJU international, 101(10), 1227-1231.
109. Parsons, J. K., Newman, V., Mohler, J. L., Pierce, J. P., Paskett, E., & Marshall, J. (2008). The Men's Eating and Living (MEAL) study:
a Cancer and Leukemia Group B pilot trial of dietary intervention for the treatment of prostate cancer. Urology, 72(3), 633-637.
110. Kotake-Nara, E., Asai, A., & Nagao, A. (2005). Neoxanthin and fucoxanthin induce apoptosis in PC-3 human prostate cancer
cells. Cancer letters, 220(1), 75-84.
111. Wu, K., Hu, F. B., Willett, W. C., & Giovannucci, E. (2006). Dietary patterns and risk of prostate cancer in US men. Cancer
Epidemiology Biomarkers & Prevention, 15(1), 167-171.
112. De Stefani, E., Ronco, A. L., Deneo-Pellegrini, H., Boffetta, P., Aune, D., Acosta, G., ... & Mendilaharsu, M. (2010). Dietary patterns
and risk of advanced prostate cancer: a principal component analysis in Uruguay. Cancer causes & control, 21(7), 1009-1016.
113. Walker, M., Aronson, K. J., King, W., Wilson, J. W., Fan, W., Heaton, J. P., ... & Morales, A. (2005). Dietary patterns and risk of
prostate cancer in Ontario, Canada. International journal of cancer, 116(4), 592-598.
114. Hoekstra, J., Hart, A., Boobis, A., Claupein, E., Cockburn, A., Hunt, A., ... & Chiodini, A. (2012). BRAFO tiered approach for benefit–
risk assessment of foods. Food and Chemical Toxicology, 50, S684-S698.
Page | 73
115. Carleton, A. J., Sievenpiper, J. L., de Souza, R., McKeown-Eyssen, G., & Jenkins, D. J. (2013). Case–control and prospective studies
of dietary α-linolenic acid intake and prostate cancer risk: a meta-analysis. BMJ open, 3(5).
116. Meng, H., Hu, W., Chen, Z., & Shen, Y. (2013). Fruit and vegetable intake and prostate cancer risk: A meta‐analysis. Asia‐Pacific
Journal of Clinical Oncology.
117. Chua, M. E., Sio, M. C. D., Sorongon, M. C., & Dy, J. S. (2012). Relationship of dietary intake of omega-3 and omega-6 Fatty acids
with risk of prostate cancer development: a meta-analysis of prospective studies and review of literature. Prostate cancer, 2012.
118. Rota, M., Scotti, L., Turati, F., Tramacere, I., Islami, F., Bellocco, R., ... & Bagnardi, V. (2012). Alcohol consumption and prostate
cancer risk: a meta-analysis of the dose–risk relation. European journal of cancer prevention, 21(4), 350-359.
119. Gilbert, R., Martin, R. M., Beynon, R., Harris, R., Savovic, J., Zuccolo, L., ... & Metcalfe, C. (2011). Associations of circulating and
dietary vitamin D with prostate cancer risk: a systematic review and dose–response meta-analysis. Cancer causes & control, 22(3),
319-340.
