Current epidemiological knowledge about the role of flavonoids in prostate carcinogenesis

Sak K.*

Summary. Numerous experimental studies have demonstrated anticancer action of polyphenolic plant metabolites. However, data about associations between dietary intake of plant-derived flavonoids and prostate cancer risk are still sparse and inconsistent. This minireview compiles the epidemiological findings published to date on the role of flavonoids in prostate tumorigenesis, discusses the reasons of inconsistencies and elicits the promising results for chemoprevention of this malignancy. Long-term consumption of high doses of soy isoflavones can be the reason of markedly lower clinically detectable prostate cancer incidence among Asian men compared to their counterparts in the Western world. The ability to metabolize daidzein to equol, the most biologically active isoflavone, by the certain intestinal bacteria also seems to contribute to this important health benefit. The increasing incidence rate of prostate cancer related to adoption of westernized lifestyle and dietary habits makes the issue of chemoprevention ever more important and directs the eyes to specific food components in the Eastern diet. If further large-scale epidemiological studies will confirm the protective effects of isoflavones against prostate cancer, this could provide an important way for prostate cancer prevention, as diet is a potentially modifiable factor in our behavioral pattern.

Submitted: March 13, 2017.
*Correspondence: E-mail: katrin.sak.001@mail.ee

The incidence rate of clinically detectable prostate cancer is historically markedly lower in Asian countries, such as Japan, China, India, Singapore or Korea, compared to the Western world, including men living in American, European countries or Canada [1–9]. This age-adjusted difference between Asian and Western populations can reach even 10- to 90-fold [10–12]. However, in parallel to adoption of westernized habits among Japanese men, the incidence rate of prostate tumors is gradually rising [1, 6, 7, 9]. Also, Japanese migrants to the United States as well as foreign-born Asian Americans reveal an increased incidence of prostate neoplasms [3, 6, 10, 11, 13–16]. On the contrary, the incidence of clinically undetectable or latent prostate cancers in autopsy studies is rather similar between men from Asian and American countries, suggesting that some components in the Asian diet can block progression of latent cancers to the more advanced form [3, 6, 8, 14]. Therefore, the etiology of prostate tumor might involve environmental factors, such as dietary habits, which can play a major role in the pathogenesis of this malignancy [1, 3–6, 11–20]. A traditional Asian diet is rich in soy-derived phytoestrogens or isoflavones that have been related to the lower incidence of prostate cancer in this region [3, 7–9, 14, 21].

Plant-based food items contain various phytochemicals, including flavonoids, with anticarcinogenic properties [18]. These polyphenolic compounds may influence prostate tumorigenesis through multiple different mechanisms; flavonoids have been shown to exhibit antioxidant, antiinflammatory, antiproliferative, proapoptotic, antiangiogenic and antimetastatic activities [13, 18, 22]. Although numerous experimental cell culture and animal studies have demonstrated different effects of these plant metabolites against prostate cancer, findings from epidemiological studies are still limited and far from definite [6, 12, 16, 19, 21, 23–26]. To comprehend the current state of epidemiological evidence on the role of various flavonoids on prostate tumorigenesis, the findings published to date about associations between dietary intake of flavonoids and the risk of prostate cancer were compiled and are analyzed in this minireview. In addition, the blood and urinary biomarker data are included in the survey. These data are presented in Tables 1–3. Also, the potential factors for inconsistencies are put forward and discussed.

Table 1. Epidemiological case-control studies on dietary intake of flavonoids and prostate cancer risk
Flavonoid
subclass
Certain
compound
Studya Population Controlsb Cases/controls Intake comparison
(low vs high, mg/day)c
Multivariate-
adjusted OR/RRd
p for trende Age, years Refe­rence
Flavonoids Italian HB 1294/1451 (Q5) 1.20 (0.92–1.58) 0.44 46–74 [21]
Flavones Italian HB 1294/1451 (Q5) 1.09 (0.85–1.40) 0.88 46–74 [21]
Flavones Apigenin American HB 83/107 0.83 (0.45–1.51) 0.54 [27]
Flavones Luteolin American HB 83/107 0.83 (0.45–1.51) 0.54 [27]
Flavonols Italian HB 1294/1451 (Q5) 1.23 (0.95–1.61) 0.26 46–74 [21]
Flavonols Kaempferol WNYDS American PB 433/538 < 1.4475 vs > 6.0568 (Q4) 0.83 (0.58–1.18) 0.80 [18]
Flavonols Kaempferol American HB 83/107 0.77 (0.42–1.43) 0.41 [27]
Flavonols Myricetin American HB 83/107 1.12 (0.61–2.06) 0.70 [27]
Flavonols Quercetin WNYDS American PB 433/538 < 10.5663 vs >25.5626 (Q4) 0.64 (0.44–0.92) 0.15 [18]
Flavonols Quercetin American HB 83/107 0.96 (0.53–1.76) 0.90 [27]
Flavanones Italian HB 1294/1451 (Q5) 0.96 (0.75–1.23) 0.56 46–74 [21]
Flavanols Italian HB 1294/1451 (Q5) 1.30 (1.01–1.69) 0.48 46–74 [21]
Isoflavones EPIC-Norfolk English PB 203/800 0.89 (0.72–1.11) 0.29 40–79 [29]
Isoflavones Italian HB 1294/1451 (Q5) 0.98 (0.76–1.26) 0.34 46–74 [21]
Isoflavones PCANDIET Scottish PB 433/483 < 0.5811 vs > 1.9828 (Q4) 1.18 (0.79–1.75) 0.87 50–74 [5]
Isoflavones CAPS Swedish PB 1499/1130 0.0008 vs 0.113 (Q4) 0.99 (0.77–1.28) 0.68 36–79 [17]
Isoflavones CAPS Swedish PB 1431/1081 0.0008 vs 0.113 (Q4) 0.99 (0.77–1.28) 0.95 35–79 [28]
Isoflavones Japanese HB 200/200 < 30.5 vs ≥ 89.9 (Q4) 0.42 (0.24–0.72) <0.01* 59–73 [1]
Isoflavones Genistein American HB 83/107 0.71 (0.39–1.30) 0.26 [27]
Isoflavones Genistein EPIC-Norfolk English PB 203/800 0.90 (0.74–1.10) 0.31 40–79 [29]
Isoflavones Genistein EPIC-Norfolk English PB 89/178 1.07 (0.84–1.37) 0.575 45–75 [12]
Isoflavones Genistein CAPS Swedish PB 1499/1130 0.00019 vs 0.0694 (Q4) 0.97 (0.75–1.26) 0.09 36–79 [17]
Isoflavones Genistein Chinese PB 133/265 < 17.9 vs > 62.0 (Q4) 0.53 (0.29–0.97) 0.058 50–89 [15]
Isoflavones Genistein Japanese HB 200/200 < 1.1 vs ≥ 2.5 (Q4) 0.58 (0.34–0.97) 0.04* 59–73 [1]
Isoflavones Daidzein American HB 83/107 0.57 (0.31–1.05) 0.07 [27]
Isoflavones Daidzein EPIC-Norfolk English PB 203/800 0.92 (0.69–1.22) 0.56 40–79 [29]
Isoflavones Daidzein EPIC-Norfolk English PB 89/178 1.06 (0.84–1.64) 0.630 45–75 [12]
Isoflavones Daidzein CAPS Swedish PB 1499/1130 0.00037 vs 0.0431 (Q4) 1.22 (0.92–1.62) 0.70 36–79 [17]
Isoflavones Daidzein Chinese PB 133/265 < 10.0 vs > 36.3 (Q4) 0.56 (0.31–1.04) 0.116 50–89 [15]
Isoflavones Daidzein Japanese HB 200/200 < 0.8 vs ≥ 1.9 (Q4) 0.55 (0.32–0.93) 0.02* 59–73 [1]
Isoflavones Biochanin A American HB 83/107 0.92 (0.50–1.70) 0.79 [27]
Isoflavones Biochanin A EPIC-Norfolk English PB 203/800 0.89 (0.69–1.14) 0.34 40–79 [29]
Isoflavones Formononetin American HB 83/107 0.99 (0.54–1.81) 0.98 [27]
Isoflavones Formononetin EPIC-Norfolk English PB 203/800 0.97 (0.83–1.13) 0.70 40–79 [29]
Isoflavones Glycitein EPIC-Norfolk English PB 203/800 0.88 (0.67–1.15) 0.34 40–79 [29]
Isoflavones Equol EPIC-Norfolk English PB 203/800 1.31 (1.00–1.71) 0.050* 40–79 [29]
Anthocyanidins Italian HB 1294/1451 1.18 (0.91–1.53) 0.22 46–74 [21]
Note: aCAPS — The Cancer Prostate Sweden Study; EPIC — The European Prospective Investigation into Cancer and Nutrition; PCANDIET — The Prostate Cancer and Diet Study; WNYDS — The Western New York Diet Study.
bHB — hospital-based; PB — population-based.
cQ4 — quartiles; Q5 — quintiles.
dOR — odds ratio; RR — relative risk.
eStatistically significant effects (p for trend < 0.05) are marked by asterisk.
Table 2. Epidemiological prospective cohort studies on dietary intake of flavonoids and prostate cancer risk
Flavonoid
subclass
Certain compound Stu­dya Population Median follow-up (years) Cases/
cohort
Intake comparison (low vs high,
mg/day)b
Multivariate-­adjusted RR/HRc p for trendd Commentse Refe­rence
Flavonoids CPS-II American 7.8 3974/
43,268
< 126.6 vs ≥ 359.5 (Q5) 1.11 (1.01–1.23) 0.02* 50–74 y; positive association for high grade cancer within the first 2 y of follow-up, suggestive inverse association after excluding these first 2 y; no associations with advanced cancer [16]
Flavonoids MEC American, Hawaiian 8.0 4404/
82,483
< 1.6 vs ≥ 7.2/ 1000 kcal (Q5) 0.93 (0.83–1.04) 0.17 45–75 y; multiethnic; no inverse association regardless of stage [23]
Flavonoids NLCS Dutch 17.3 3362/
58,279
13.1 vs 40.6 (Q4) 1.00 (0.84–1.18) 0.74 55–69 y; no significant effect modification by cancer stage [33]
Flavonoids FMC Finnish 24.0 62/9959 < 2.1 vs > 4.8 (Q4) 1.39 (0.56–3.46) 15–99 y [30]
Flavonoids FMC Finnish 30.0 95/5218 4.3 vs 26.9 (Q4) 1.11 (0.61–2.01) 0.57 [31]
Flavonoids KIHD Finnish 16.2 138/2590 9.1 vs 416.3 (Q4) 1.16 (0.58–2.34) 0.831 42–60 y [32]
Flavones CPS-II American 7.8 3974/43,268 < 0.5 vs ≥ 2.2 (Q5) 1.05 (0.95–1.16) 0.19 50–74 y; no associations for high grade or advanced cancers within or without the first 2 y of follow-up [16]
Flavones KIHD Finnish 16.2 138/2590 (Q4) 0.71 (0.30–1.65) 0.561 42–60 y [32]
Flavonols CPS-II American 7.8 3974/
43,268
< 8.9 vs ≥ 20.5 (Q5) 1.10 (0.99–1.21) 0.05* 50–74 y; suggestive positive association for high grade cancer within the first 2 y of follow-up, not after excluding these first 2 y; no associations with advanced cancer [16]
Flavonols KIHD Finnish 16.2 138/2590 (Q4) 1.53 (0.72–3.23) 0.585 42–60 y [32]
Flavonols Kaempferol NLCS Dutch 17.3 3362/
58,279
2.5 vs 11.2 (Q4) 0.98 (0.83–1.16) 0.73 55–69 y; significant inverse association for stage IV cancer [0.78 (0.61–1.00), p = 0.03*], not for non-advanced cancer [33]
Flavonols Kaempferol FMC Finnish 30.0 95/5218 0.1 vs 0.8 (Q4) 1.03 (0.53–2.02) 0.54 [31]
Flavonols Myricetin NLCS Dutch 17.3 3362/
58,279
0.4 vs 2.6 (Q4) 0.89 (0.75–1.05) 0.41 55–69 y; significant inverse association for stage IV cancer [0.71 (0.55–0.91), p = 0.03*], not for non-advanced cancer [33]
Flavonols Myricetin FMC Finnish 30.0 95/5218 0 vs 0.11 (Q4) 0.43 (0.22–0.86) 0.002* [31]
Flavonols Quercetin NLCS Dutch 17.3 3362/
58,279
8.8 vs 28.5 (Q4) 1.01 (0.85–1.19) 0.98 55–69 y; no significant effect modification by cancer stage [33]
Flavonols Quercetin FMC Finnish 30.0 95/5218 1.5 vs 3.9 (Q4) 0.76 (0.40–1.42) 0.35 [31]
Flavanones CPS-II American 7.8 3974/
43,268
< 7.5 vs ≥ 38.2 (Q5) 1.08 (0.98–1.20) 0.31 50–74 y; positive association for high grade cancer within the first 2 y of follow-up, not after excluding these first 2 y; no associations with advanced cancer [16]
Flavanones KIHD Finnish 16.2 138/2590 (Q4) 0.90 (0.37–2.20) 0.518 42–60 y [32]
Flavanones Hesperetin FMC Finnish 30.0 95/5218 0 vs 15.4 (Q4) 1.47 (0.80–2.71) 0.26 [31]
Flavanones Naringenin FMC Finnish 30.0 95/5218 0 vs 4.7 (Q4) 1.48 (0.80–2.73) 0.27 [31]
Flavanols CPS-II American 7.8 3974/
43,268
< 10.4 vs ≥ 37.9 (Q5) 1.18 (1.06–1.30) 0.01* 50–74 y; suggestive positive association for high grade cancer within the first 2 y of follow-up, not after excluding these first 2 y; no associations with advanced cancer [16]
Flavanols NLCS Dutch 17.3 3362/
58,279
14.5 vs 98.7 (Q4) 0.97 (0.82–1.15) 0.58 55–69 y; significant inverse association for stage IV cancer [0.73 (0.57–0.95), p = 0.01*], not for non-advanced cancer [33]
Flavanols KIHD Finnish 16.2 138/2590 (Q4) 1.37 (0.65–2.89) 0.82 42–60 y [32]
Flavanols Catechin NLCS Dutch 17.3 3362/
58,279
1.9 vs 6.8 (Q4) 1.03 (0.88–1.22) 0.83 55–69 y; no significant effect modification by cancer stage [33]
Flavanols Epicatechin NLCS Dutch 17.3 3362/
58,279
5.3 vs 20.8 (Q4) 1.01 (0.86–1.20) 0.62 55–69 y; significant inverse association for stage III/IV [0.84 (0.68–1.04), p = 0.05*] and IV cancer [0.74 (0.57–0.95), p = 0.01*], not for non-advanced cancer [33]
Isoflavones CPS-II American 7.8 3974/
43,268
< 0.029 vs ≥ 0.144 (Q5) 1.11 (1.01–1.22) <0.001* 50–74 y; no associations for high grade or advanced cancers within or without the first 2 y of follow-up [16]
Isoflavones Genistein MEC American, Hawaiian 8.0 4404/
82,483
< 0.7 vs ≥ 3.1/ 1000 kcal (Q5) 0.94 (0.84–1.04) 0.16 45–75 y; multiethnic; no inverse association regardless of stage [23]
Isoflavones Genistein JPHC Japanese 5.0 307/
43,509
< 13.2 vs ≥ 32.8 (Q4) 0.71 (0.48–1.03) 0.22 45–74 y; non-significant inverse association for localized, not advanced cancer; significant inverse association for men > 60 y with localized cancer [0.52 (0.30–0.90), p =0.03*], not advanced cancer or for younger men [3]
Isoflavones Daidzein MEC American, Hawaiian 8.0 4404/
82,483
< 0.7 vs ≥ 3.2/1000 kcal (Q5) 0.92 (0.82–1.02) 0.09 45–75 y; multiethnic; no inverse association regardless of stage [23]
Isoflavones Daidzein JPHC Japanese 5.0 307/
43,509
< 8.5 vs ≥ 20.4 (Q4) 0.77 (0.52–1.13) 0.43 45–74 y; non-significant inverse association for localized, not advanced cancer; significant inverse association for men > 60 y with localized cancer [0.50 (0.28–0.88), p = 0.04*], not advanced cancer or for younger men [3]
Isoflavones Glycitein MEC American, Hawaiian 8.0 4404/
82,483
< 0.18 vs ≥ 0.80/1000 kcal (Q5) 0.91 (0.82–1.01) 0.07 45–75 y; multiethnic; no inverse association regardless of stage [23]
Anthocyanidins CPS-II American 7.8 3974/
43,268
< 5.9 vs ≥ 18.0 (Q5) 1.10 (1.00–1.22) 0.13 50–74 y; no associations for high grade or advanced cancers within or without the first 2 y of follow-up [16]
Anthocyanidins KIHD Finnish 16.2 138/2590 (Q4) 0.59 (0.24–1.41) 0.974 42–60 y [32]
Note: aCPS-II — The Cancer Prevention Study II Nutrition Cohort; FMC — The Finnish Mobile Clinic Health Examination Survey; JPHC — The Japan Public Health Center-based prospective study; KIHD — The Kuopio Ischaemic Heart Disease Risk Factor Study; MEC — The Multiethnic Cohort Study; NLCS — The Netherlands Cohort Study.
bQ4 — quartiles; Q5 — quintiles.
cRR — relative risk; HR — hazard ratio.
dStatistically significant effects (p for trend < 0.05) are marked by asterisk.
eAge in baseline.
Table 3. Epidemiological studies on biomarkers of flavonoids and prostate cancer risk
Flavonoid
subclass
Certain compound Bio-marker Studya Population Cases/ controls Multivariate-­adjusted OR/RRb p for trendc Commentsd Refe­rence
Isoflavones Serum EPIC-Norfolk English 191/815 1.01 (0.93–1.10) 0.809 45–75 y [34]
Isoflavones Serum PCANDIET Scottish 238/198 1.24 (0.69–2.20) 0.64 50–74 y [5]
Isoflavones Urinary EPIC-Norfolk English 152/665 0.98 (0.90–1.08) 0.727 45–75 y [34]
Isoflavones Genistein Plasma EPIC European 1605/1697 1.00 (0.79–1.27) 0.82 43–76 y; from eight countries; no associations as stratified by cancer stage (localized or advanced) or histological grade (low or high), or age of men at diagnosis (< 60 y or ≥ 60 y) [35]
Isoflavones Genistein Plasma EPIC European 950/1042 0.74 (0.54–1.00) 0.051* 43–76 y; from eight countries [24]
Isoflavones Genistein Plasma JPHC Japanese 201/402 0.66 (0.40–1.08) 0.08 40–69 y; significant inverse association for localized cancer [0.54 (0.29–1.01), p = 0.03*], not advanced cancer [5]
Isoflavones Genistein Serum JACC Japanese 52/151 0.76 (0.32–1.82) 0.54 ≥ 40 y; suggestive inverse association for men with T ≥ 6.9 nM, or T ≥ 6.9 nM and PSA ≤ 10 ng/ml for cases, < 4 ng/ml for controls [25]
Isoflavones Genistein Plasma Chinese 46/54 0.31 (0.13–0.71) 0.006* [4]
Isoflavones Genistein Serum EPIC-Norfolk English 191/815 0.99 (0.93–1.05) 0.78 45–75 y [34]
Isoflavones Genistein Serum EPIC-Norfolk English 89/178 1.01 (0.83–1.23) 0.938 45–75 y [12]
Isoflavones Genistein Serum PCANDIET Scottish 238/198 1.36 (0.76–2.43) 0.37 50–74 y [5]
Isoflavones Genistein Urinary MEC American, Hawaiian 249/404 0.72 (0.40–1.31) 0.09 45–75 y; multiethnic; suggestive inverse association for localized and advanced cancer, and for Latinos and Whites (not African Americans and Japanese Americans) [37]
Isoflavones Genistein Urinary Jamaican 175/194 1.23 (0.67–2.56) 0.502 40–80 y; no associations as stratified by cancer grades (low or high) [36]
Isoflavones Genistein Urinary EPIC-Norfolk English 152/665 1.00 (0.95–1.05) 0.86 45–75 y [34]
Isoflavones Genistein Urinary EPIC-Norfolk English 89/178 0.92 (0.76–1.12) 0.415 45–75 y [12]
Isoflavones Daidzein Plasma EPIC European 950/1042 0.80 (0.60–1.07) 0.209 43–76 y; from eight countries [24]
Isoflavones Daidzein Plasma JPHC Japanese 201/402 0.78 (0.49–1.25) 0.44 40–69 y; suggestive inverse association for localized, not advanced cancer [6]
Isoflavones Daidzein Serum EPIC-Norfolk English 89/178 1.01 (0.86–1.18) 0.902 45–75 y [12]
Isoflavones Daidzein Serum EPIC-Norfolk English 191/815 0.99 (0.93–1.05) 0.68 45–75 y [34]
Isoflavones Daidzein Serum PCANDIET Scottish 247/200 1.34 (0.76–2.38) 0.21 50–74 y [5]
Isoflavones Daidzein Serum JACC Japanese 52/151 0.74 (0.31–1.76) 0.50 ≥40 y; suggestive inverse association for men with T ≥ 6.9 nM, or T ≥ 6.9 nM and PSA ≤ 10 ng/ml for cases, < 4 ng/ml for controls [25]
Isoflavones Daidzein Urinary MEC American, Hawaiian 249/404 0.55 (0.31–0.98) 0.03* 45–75 y; multiethnic; suggestive inverse associations for localized and advanced cancer, and for different ethnic groups (African Americans, Japanese Americans, Latinos and Whites) [37]
Isoflavones Daidzein Urinary Jamaican 175/194 0.85 (0.47–1.54) 0.600 40–80 y; no associations as stratified by cancer grades (low or high) [36]
Isoflavones Daidzein Urinary EPIC-Norfolk English 152/665 0.97 (0.92–1.03) 0.31 45–75 y [34]
Isoflavones Daidzein Urinary EPIC-Norfolk English 89/178 0.99 (0.83–1.17) 0.866 45–75 y [12]
Isoflavones Glycitein Plasma JPHC Japanese 201/402 0.78 (0.48–1.26) 0.51 40–69 y; no associations as stratified by cancer stage (localized or advanced) [6]
Isoflavones Glycitein Serum EPIC-Norfolk English 191/815 1.01 (0.95–1.08) 0.67 45–75 y [34]
Isoflavones Glycitein Serum EPIC-Norfolk English 89/178 1.08 (0.49–2.38) 0.848 45–75 y [12]
Isoflavones Glycitein Urinary EPIC-Norfolk English 152/665 1.01 (0.96–1.06) 0.81 45–75 y [34]
Isoflavones Glycitein Urinary EPIC-Norfolk English 89/178 1.01 (0.82–1.26) 0.901 45–75 y [12]
Isoflavones O-Desmethylangolensin Serum EPIC-Norfolk English 191/815 0.97 (0.91–1.03) 0.33 45–75 y [34]
Isoflavones O-Desmethylangolensin Serum EPIC-Norfolk English 89/178 0.68 (0.36–1.29) 0.236 45–75 y [12]
Isoflavones O-Desmethylangolensin Urinary EPIC-Norfolk English 152/665 0.98 (0.94–1.03) 0.42 45–75 y [34]
Isoflavones O-Desmethylangolensin Urinary EPIC-Norfolk English 89/178 0.93 (0.73–1.18) 0.540 45–75 y [12]
Isoflavones Equol Plasma EPIC European 950/1042 0.99 (0.70–1.39) 0.461 43–76 y; from eight countries [24]
Isoflavones Equol Plasma JPHC Japanese 201/402 0.60 (0.36–0.99) 0.04* 40–69 y; significant inverse association for localized cancer [0.43 (0.22–0.82), p = 0.02*], not advanced cancer [6]
Isoflavones Equol Serum EPIC-Norfolk English 191/815 1.02 (0.96–1.08) 0.52 45–75 y [34]
Isoflavones Equol Serum EPIC-Norfolk English 89/178 1.32 (0.73–2.38) 0.358 45–75 y [12]
Isoflavones Equol Serum PCANDIET Scottish 247/200 1.07 (0.71–1.61) 0.75 50–74 y [5]
Isoflavones Equol Serum JACC Japanese 52/151 0.39 (0.15–0.98) 0.046* ≥40 y; suggestive inverse association for men with T ≥ 6.9 nM, or T ≥ 6.9 nM and PSA ≤ 10 ng/ml for cases, < 4 ng/ml for controls [25]
Isoflavones Equol Urinary MEC American, Hawaiian 249/404 1.32 (0.84–2.08) 0.08 45–75 y; multiethnic; no associations for localized and advanced cancer or different ethnic groups (African Americans, Japanese Americans, Latinos and Whites) [37]
Isoflavones Equol Urinary Jamaican 175/194 0.48 (0.26–0.87) 0.020* 40–80 y; significant inverse association for high grade cancer [0.29 (0.13–0.60), p = 0.001*], not low grade cancer [36]
Isoflavones Equol Urinary EPIC-Norfolk English 152/665 1.03 (0.97–1.07) 0.20 45–75 y [34]
Isoflavones Equol Urinary EPIC-Norfolk English 89/178 1.21 (0.93–1.57) 0.147 45–75 y [12]
Note: aEPIC — The European Prospective Investigation into Cancer and Nutrition; JACC — The Japan Collaborative Cohort Study; JPHC — The Japan Public Health Center-based Prospective Study; MEC — The Multiethnic Cohort Study; PCANDIET — The Prostate Cancer and Diet Study.
bOR — odds ratio; RR — relative risk; the risk for equol is presented in comparison to equol non-producers.
cStatistically significant effects (p for trend < 0.05) are marked by asterisk.
dPSA — prostate-specific antigen; T — testosterone.

Epidemiological data about the role of flavonoids on prostate cancer risk

Case-control studies

No statistically significant associations with the risk of prostate cancer were found for total flavonoids or flavonoid subclasses, including flavones, flavonols, flavanones, flavanols, isoflavones and anthocyanidins in Italian population, showing no protective effects of flavonoids against prostate tumorigenesis [21]. These results were in general agreement with two other case-control studies, revealing no significant relationships between intake of flavones (apigenin, luteolin) and flavonols (kaempferol, myricetin) and prostate cancer incidence among American men [18, 27]. However, a suggestive inverse association was still reported for flavonol quercetin with a 36% decrease in prostate cancer risk for men in the highest quartile of quercetin consumption compared to those in the lowest quartile of intake [18].

Similarly to the work with Italian men, no overall relationships between consumption of total or individual isoflavones (genistein, daidzein) and prostate cancer risk were observed also in Swedish [17, 28], Scottish [5], or English population [12, 29]. However, these results are different from the findings of case-control studies conducted with Japanese [1] and Chinese men [15], which observed a protective effect of intake of both genistein and daidzein against prostate cancer. At that, Nagata et al. described a significant 42% and 45% reduction of prostate cancer risk with high dietary consumption of genistein and daidzein, respectively [1], and Lee et al. reported a 47% and 44% decrease in the tumor risk with high consumption of these isoflavones, respectively [15]. These results demonstrate that isoflavones can indeed contribute to a protective effect against prostate neoplasms. Also, Strom et al. observed some suggestive, still non-significant, lowering in prostate cancer incidence with high dietary consumption of daidzein and genistein among American men [27]. As the daily intake of soy foods in European countries is very low compared to Asian men (this difference can be even about hundredfold), it is possible that these levels were too low for a preventive effect against prostate cancer to be become obvious [12, 21] (see Table 1).

Prospective cohort studies

Using the prospective cohort study design, no significant inverse associations with prostate cancer risk were found for intake of total flavonoids as well as flavonoid subclasses flavones, flavanones and anthocyanidins in different populations, including American, Hawaiian, Dutch, and Finnish men [23, 30–33]. Although Wang et al. observed a weak positive relationship with intake of total flavonoids, flavanols or isoflavones, but also flavonols or flavanones with overall prostate cancer incidence, the relative risk varied over follow-up time and the positive association disappeared when the cases diagnosed during the first two follow-up years were excluded from the analysis [16]. The authors proposed that such pattern could be affected by behaviors associated with prostate cancer screening [16]. For flavonols, Knekt et al. found a statistically significant 57% decrease in prostate cancer risk with high dietary intake of myricetin among Finnish men; but no risk modification by other common flavonols like kaempferol or quercetin [31]. Geybels et al. conducted a prospective cohort study with Dutch men and stratification of the analysis by cancer stage revealed a significant inverse association between intake of both kaempferol and myricetin (not quercetin) with stage IV prostate cancer and not with overall or non-advanced neoplasms (stage I/II tumors) [33]. Similar trend became evident also for total flavanols and epicatechin (not catechin), as the significant inverse association was found for advanced stage prostate tumor and not non-advanced cancer [33].

Although both Park et al. and Kurahashi et al. reported no significant relationships between intake of individual isoflavones (genistein, daidzein) and total prostate cancer risk [3, 23], Kurahashi et al. still found an inverse association between intake of both genistein and daidzein and localized cancer [3]. In particular, Japanese men older than 60 years had about 50% lower risk of localized prostate cancer with high dietary consumption of genistein and daidzein [3] (see Table 2). Thus, compared to findings from case-control studies, the evidence from prospective cohort studies about the protective effects of flavonoids against prostate cancer is much more sparse being restricted to certain subgroups (stratified by age of subjects or cancer stage).

Biomarker studies

Unlike the works assessing retrospectively or prospectively dietary intake of different flavonoids, biomarker studies have focused only to isoflavones and their metabolites. At that, no associations were observed between total serum or urinary isoflavones and prostate cancer risk in English [34] or Scottish men [5]. Null associations were detected also for individual serum isoflavones and their urinary biomarkers (genistein, daidzein, glycitein, O-desmethylangolensin, equol) within English men [12, 34] and for serum isoflavones among older Scottish men [5]. In a large prospective investigation performed with men from eight European countries, no significant relationship was reported for plasma genistein, daidzein or equol and prostate cancer risk [24, 35]. In the work conducted with Jamaican men, no association was found between urinary excretion of genistein or daidzein and prostate cancer incidence, but a significant inverse relationship was found for urinary equol concentration when non-producers of equol were assessed as reference group [36]. This study indicated that men who were able to produce equol from daidzein by the aid of specific intestinal microflora were at reduced risk of total and high grade (but not low grade) prostate cancer as compared to non-producers of this metabolite [36]. However, another study that measured urinary biomarkers of isoflavones did not observe such protective effect for equol [37]. Quite contrary, urinary excretion of daidzein was significantly inversely associated with prostate cancer risk among multiethnic men living in Hawaii and Los Angeles and a respective suggestive trend was indicated also for genistein, regardless of the stage of tumor [37].

Differently from works performed with European and American men, their Asian counterparts consume substantially higher amounts of soy foods and consequently also isoflavones. Wu et al. reported a 69% reduced prostate cancer risk among Chinese men consuming high amounts of genistein as evaluated by plasma genistein level [4]. Kurahashi et al. described a significant 44% reduction in the risk of localized (not advanced) prostate tumor among Japanese men with higher plasma genistein concentrations and a similar suggestive tendency was reported also for plasma daidzein level [6]. Plasma equol level was inversely related to both overall as well as localized prostate cancer incidence, not modifying the risk of advanced cancer [6]. A significant 61% decrease in prostate cancer risk was detected also in another study with Japanese men with high serum equol levels and similar suggestive, still non-significant, trend was observed for genistein and daidzein when the subjects with low serum testosterone levels were excluded [25]. These studies with Asian men support that high doses of isoflavones can be indeed preventive against prostate cancer development (see Table 3).

Potential factors giving rise to inconcistencies

As come into sight from the epidemiological findings presented in above paragraphs, there are still relatively few data published about the associations between flavonoids and prostate cancer risk, and these results are rather inconclusive. There are seve­ral potential reasons providing some explanations to these inconsistencies.

First, assessment of intake of flavonoids has been limited for a long time due to lack of data about food composition. Exposure to flavonoids can be estimated by either food diaries or food frequency questionnaires and these data are further translated to dietary intake using quantitative food composition databases [12]. However, reliable information about the content of flavonoids in different plant-derived food items was not available until the last decade, when several databases were made public for scientific community [16, 21, 23]. However, these food composition databases are not able to capture the variability of flavonoid content attributable to cultivars, growing methods and environmental conditions, exposure to sunlight and heat, or harvesting time and processing practices [21, 33, 38]. Also, these data tables can not detect all sources of flavonoids in the human diet [37]. Furthermore, some questions arise by adaptation of food composition databases created in the US to the diet consumed in European countries [5, 21].

Second, another possible reason for the inconsistent results of epidemiological studies can come from the differences in level of flavonoids intake. For example, it has been demonstrated that the mean daily consumption of isoflavones is low in Western populations (< 3 mg), whereas it is considerably higher among Asian subjects reaching even 70–100 mg and being for most Asian men in the range of 30 to 50 mg [1, 4, 5, 7, 15–17, 23]. Therefore, it can be that the consumption of isoflavones in Western countries is too low to observe a protective action against prostate tumor [4, 5, 9, 10, 12, 16, 23, 35]. Whether similar effects exist also for other flavonoids and their subclasses, i.e. the risk-lowering response manifests only at a certain level, is still unknown. The possible age periods of the highest susceptibility to beneficial effects of flavonoids have also remained to be determined. Furthermore, we can not rule out the possibility that follow-up times of prospective cohort studies were not long enough for detection of protective action, especially in the case of advanced neoplasms [16].

Third, it is well known that following to their ingestion, flavonoids undergo an extensive metabolic conversion in the intestine and liver, as a result of which various conjugates with potentially altered biological activity enter bloodstream and reach target tissues, and are finally excreted in urine [11, 12]. Blood and urinary biomarkers can reflect dietary consumption of flavonoids, capturing intake from all food sources, considering also their metabolic biotransformation [24, 34]. Therefore, direct measurement of these biomarkers might provide a more relevant insight into the associations between exposure to flavonoids and prostate cancer risk than assessment of food consumption [34]. However, besides dietary intake, the levels of flavonoids in biological fluids are affected by intestinal microflora, diet composition and content of fat in the food, consumption of alcohol, and the use of antibiotics as well as possible bowel diseases [5, 28, 36]. Moreover, biomarkers reflect only short-term intake of flavonoids due to their brief half-lives in the human body and might not accurately reflect the long-term dietary habits; especially, when only one serum measurement or spot urine test is used [2, 5, 6, 35–37]. Thus, estimation of dietary intake has remained indispensable in epidemiological studies, whereas measurement of biomarkers can complement dietary assessment with bioavailability issues. At that, about 30–50% of people are able to metabolize daidzein to equol, the most biologically active isoflavone, by the aid of special intestinal bacteria and this ability varies within different ethnic populations [8, 9, 19, 36, 38]. Moreover, it has been shown that the prevalence of equol-producers is generally lower among subjects suffering from prostate cancer compared to healthy controls [2, 8, 19]. For example, it was shown that just 29% of Japanese men with prostate cancer could produce equol, compared to 46% of controls and these estimates were 30% and 59% for Korean cases and controls, respectively [8]. If further studies confirm that equol can lower the prostate cancer risk and intestinal environment without equol-converting bacteria is indeed a risk factor for this malignancy, a possible prevention strategy for future would be to improve intestinal microflora by probiotic technology, enabling it to generate equol [7, 9].

In the case of retrospective study design, the recall bias is unavoidable. Patients could recall their food consumption differently than controls and it is also possible that they changed their diet after cancer was diagnosed [15, 21]. However, this misclassification was probably non-differential, as the knowledge about possible role of flavonoids on prostate carcinogenesis was not widespread among general public [15, 28]. Analysis of data collected prospectively might reduce the problems of disease-related influences on dietary intake of flavonoid-rich food items or their biomarker levels [34, 37]. However, the number of cases could not be large enough for acceptably strong statistical power, especially for subgroup analyses stratified by different parameters [19, 33, 37].

Another common concern in epidemiological studies is potential confounding. Despite adjustment for diverse aspects, it is always possible that residual confounders due to uncontrolled or as yet unknown factors might remain [5, 12, 15, 16, 26, 27, 36]. For example, most of the studies did not collect information about testing of prostate specific antigen that might be a confounder [16, 33]. Also, only few investigations considered histological grade or tumor stage, although risk factors of prostate cancer can vary within subgroups differentiated by di­sease aggressiveness or stage [8, 26, 33]. In addition, lifestyle related to consumption of higher doses of flavonoids or some dietary covariates might contribute to the susceptibility to prostate carcinogenesis [3, 5, 10, 16]. Furthermore, Hedelin et al. indicated that health benefits of isoflavone intake against prostate tumor can be restricted to men carrying specific genetic variant in estrogen receptor β gene [28].

Conclusions and further perspectives

High incidence rate, slow progression and long course make prostate cancer a perfect disorder for chemoprevention which onset, development and progression can be influenced by certain dietary constituents [20]. There are several epidemiological findings suggesting that high intake of specific flavonoids can lead to reduced prostate cancer risk. Indeed, consumption of isoflavones at doses characteristic to Asian population may modulate the risk of prostate cancer and therefore, high intake of these phytoestrogens might be an efficient protective means against prostate tumorigenesis among Chinese and Japanese men. The stronger preventive effect of isoflavones on localized tumor compared to advanced neoplasm is in agreement with the expression of estrogen receptor β in normal prostate tissue and early stages neoplasm, decline in expression during malignant transformation and complete loss in advanced tumor with higher metastatic potential and mortality; supposing that preventive effects of isoflavones are mediated via this receptor subtype [3, 6, 28]. Based on the autopsy studies the rate of latent tumors is similar in Asian and Western countries and the significantly lower incidence of clinically detectable prostate tumors in Asian men can come from blocking and delaying the latent cancer progression by intake of high amounts of isoflavones [3, 6]. An important factor in health benefits of high consumption of isoflavones seems to arise also from the ability to metabolize daidzein to equol by possessing a specific intestinal bacterial combination. Whether ethnic variation in this metabolic conversion may provide some explanation to the differences in prostate cancer incidence remains to be further clarified. On the other hand, some epidemiological studies with Western men have demonstrated the protective action of certain flavonols or flavanols manifesting particularly against advanced, later stage tumor. As a whole, these data indicate that through intake of various plant-derived phytochemicals nature can provide protection and chemoprevention against different stages of prostate cancer.

There is no doubt that large-scale and preferentially prospectively designed studies are needed to further investigate the association between exposure to flavonoids and the risk of prostate cancer. At that, subjects with consumption of wider concentration range of these plant metabolites as well as stratification by different factors, including disease stage, are highly needed. If the protective effects will be confirmed, chemopreventive intervention strategies via modification of dietary habits can be developed to reduce the prostate cancer incidence, especially among at-risk men. Last but not least, although adverse effects of flavonoids on the prostate tissue have not been reported, studies considering safety profile of these phytochemicals are never too much.

Conflict of interest

None declared.

REFERENCES

No Comments » Add comments
Leave a comment

ERROR: si-captcha.php plugin says GD image support not detected in PHP!

Contact your web host and ask them why GD image support is not enabled for PHP.

ERROR: si-captcha.php plugin says imagepng function not detected in PHP!

Contact your web host and ask them why imagepng function is not enabled for PHP.