Abstract
Background
The role of lycopene in prostate cancer prevention remains controversial. We examined the associations between dietary lycopene intake and prostate cancer, paying particular attention to the influence of prostate-specific antigen screening, and evaluated tissue biomarkers in prostate cancers in relation to lycopene intake.
Methods
Among 49898 male health professionals, we obtained dietary information through questionnaires and ascertained total and lethal prostate cancer cases from 1986 through January 31, 2010. Cox regression was used to estimate multivariable hazard ratios (HRs) and 95% confidence intervals (CIs). Tissue microarrays and immunohistochemistry were used to assess tumor biomarker expression in a subset of men. Two-sided χ2 tests were used to calculate the P values.
Results
Higher lycopene intake was inversely associated with total prostate cancer and more strongly with lethal prostate cancer (top vs bottom quintile: HR = 0.72; 95% CI = 0.56 to 0.94; P trend = .04). In a restricted population of screened participants, the inverse associations became markedly stronger (for lethal prostate cancer: HR = 0.47; 95% CI = 0.29 to 0.75; P trend = .009). Comparing different measures of dietary lycopene, early intake, but not recent intake, was inversely associated with prostate cancer. Higher lycopene intake was associated with biomarkers in the cancer indicative of less angiogenic potential.
Conclusions
Dietary intake of lycopene was associated with reduced risk of lethal prostate cancer and with a lesser degree of angiogenesis in the tumor. Because angiogenesis is a strong progression factor, an endpoint of lethal prostate cancer may be more relevant than an endpoint of indolent prostate cancer for lycopene in the era of highly prevalent prostate-specific antigen screening.
In spite of recent controversy regarding prostate cancer screening, overdiagnosis, and excessive therapy, prostate cancer remains the second leading cause of cancer mortality in American men, accounting for more than 30000 deaths (1).
Human prostatic cancer lesions can arise in middle age and often remain latent years before the development of any symptoms (2). Before the widespread use of the prostate-specific antigen (PSA) test for screening, most of the cancers diagnosed were palpable or caused symptoms and were usually detected at more-advanced stages. However, most cancers diagnosed in the PSA era are found at localized stages, are asymptomatic and biologically indolent, and would have remained undiagnosed in the pre-PSA era. Because risk factors for prostate cancer may differ by subtype of cancer and may act at different stages of the disease (3), epidemiological studies of prostate cancer conducted in different settings may produce different results.
Lycopene, a carotenoid with multiple bioactivities, is found abundantly in tomato, tomato-based products, pink grapefruit, and watermelon (4). A number of studies have investigated lycopene, or lycopene-rich food such as tomato and tomato-based products, in relation to prostate cancer risk (5–25). In a meta-analysis of studies published up to 2003 (26), high intakes of tomato or tomato-based products was associated with a 10% to 20% reduction in prostate cancer risk, and high serum or plasma concentrations of lycopene were associated with a 25% reduced risk. Among recent studies of lycopene and prostate cancer, some support an inverse association (18,19,27,28), whereas others present null findings (8,9,17,20,22,29). The heterogeneity of prostate cancers diagnosed in the PSA era may contribute to this inconsistency.
Previously, we reported that dietary intake of lycopene was associated with a 20% lower risk incident of prostate cancer in a prospective study of health professionals (5). This inverse association persisted in an updated 2002 analysis (10). In the same population, high level of lycopene in the plasma was associated with a non-statistically significant 34% reduction in risk of overall prostate cancer (27).
In this study, we aimed to reassess the association between prostate cancer and lycopene intake based on dietary data from 1986 to 2006 and follow-up through 2010 within this same cohort, focusing on four specific aims. First, in addition to total cancer incidence, we investigated the lycopene intake in relation to risk of lethal prostate cancer. Second, we identified a restricted population of participants who had been initially screened negative by PSA testing at baseline to better assess the association of lycopene with incident prostate cancer in contrast with indolent prevalent prostate cancers discovered during the early part of the PSA era with initial screening tests. Third, because multiple dietary assessments of lycopene intake were available, we examined whether diet at different time periods influenced the results. Finally, to identify the potential anticancer mechanism of lycopene, we evaluated tumor tissue biomarkers of angiogenesis, apoptosis, and cell proliferation and differentiation in relation to lycopene intake.
Methods
Study Population
The Health Professionals Follow-up Study (HPFS) is an ongoing prospective cohort study that consists of 51529 US male health professionals (dentists, optometrists, osteopaths, podiatrists, pharmacists, veterinarians) aged between 40 and 75 years at baseline in 1986. The men were followed through questionnaires every 2 years to obtain information on lifestyle factors and health outcomes. A food frequency questionnaire (FFQ) was mailed to assess usual diet every 4 years. Participants who were diagnosed with cancer before 1986, those without a completed FFQ at baseline, those who reported implausibly high (>4200 calories) or low (<800 calories) energy intake, or those who had left more than 70 items blank on the FFQ were excluded from the baseline cohort. The HPFS is approved by the Human Subjects Committee at the Harvard School of Public Health, and written informed consents were provided by all participants.
Assessment of Dietary Intake
Dietary intake was assessed by self-administered semiquantitative FFQ in 1986, 1990, 1994, 1998, 2002, and 2006. The FFQ contains a list of 131 food and beverage items for which commonly used units or portion sizes are specified. The participants were asked to report how often, on average, in the previous year, they consumed each food item. Nutrient intakes were then calculated by multiplying the frequency of consuming a certain food or beverage item by the nutrient content of that serving and then summing contributions from all food and beverage items using composition values from US Department of Agriculture (USDA) sources supplemented with other data. Carotenoid values in our nutrient database were updated using the USDA–National Cancer Institute database (30,31). The carotenoid content of tomato-based products was further updated with values from the USDA, which were determined through reversed-phase high-performance liquid chromatography (32). In the 1986 FFQ, food sources for lycopene included tomato, tomato sauce, tomato juice, pizza, watermelon, and pink grapefruit. The subsequent FFQs have also included salsa, picante or taco sauce, and ketchup or red chili sauce. Among men for whom blood sample measurement of lycopene was available (n = 1200), the mean plasma lycopene levels were 681.7, 761.6, 820.0, 884.8, and 934.5mol/L for the lowest to the highest quintile of dietary lycopene intake, respectively.
Ascertainment and Classification of Prostate Cancer
Prostate cancer diagnoses were initially identified through self-report by the participants or their next of kin on the biennial questionnaire, then confirmed by review of medical records and pathology reports. Clinical information, such as tumor stage and PSA at diagnosis and Gleason score, was acquired through a standardized review of medical records. Deaths were ascertained through repeated mailings, telephone calls to nonrespondents, and searches of the National Death Index. All causes of death were confirmed by extensive review of death certificates and medical records. Follow-up rates for cancer and for mortality were greater than 96% and nearly 100%, respectively. Participants who had prostate cancer were separately followed through an annual prostate cancer–specific questionnaire. Detailed information on treatment and development of metastasis was obtained by questionnaires and collection of medical records.
In this study, we considered total prostate cancer as all incident cases during the follow-up excluding stage T1a cancers, which are discovered incidentally during treatment for benign prostatic hypertrophy. Lethal prostate cancers were defined as cancers that caused death or distant metastases before the end of follow-up. Organ-confined prostate tumors are those that were confined within prostate gland or had limited extraprostatic extension (stage T1b, T1c, T2, T3a and N0 or Nx and M0 at diagnosis) (33). Through 2010, there were 5728 total incident prostate cancers and 658 lethal prostate cancers.
Archival formalin-fixed, paraffin-embedded tumor specimens have been retrieved for a subset of men with clinically localized prostate cancer diagnosed through 2002 (n = 1180) who underwent prostatectomy (95%) and transurethral resection of the prostate (5%), as previously described (34,35).
Immunohistochemistry
Angiogenesis.
Evaluation of angiogenic biomarkers has been described in detail previously (34). Briefly, protein expression of endothelial cell marker CD34 was ascertained on 5-micron sections. The size and architecture of microvessels were quantified by semi-automated image analysis using Image ProPlus 4.5 software (Media Cybernetics, Rockville, MD). Vessel size was determined as the average area comprised by a vessel (μm2). The irregularity of the vessel lumen was calculated by perimeter2/4 × Π × area, with a value of 1.0 indicating a perfect circle and values greater than 1.0 indicating increasing irregularity.
Tissue Microarrays.
The construction and the immunohistochemical staining of tissue microarrays have been described previously (36,37). Briefly, high-density tumor tissue microarrays were constructed by sampling 0.6-mm paraffin-embedded tumor tissue cores (at least three cores per subject) from men with localized prostate cancer. Our pathologists (M. Loda, M. Fiorentino, S. Finn, R. Flavin) reviewed hematoxylin and eosin slides to confirm prostate cancer and to provide uniform Gleason grading.
Apoptosis.
The TUNEL assay was performed on 5-micron sections of the tumor tissue microarrays to identify the percentage of tumor cells undergoing apoptosis. The procedure was carried out using the Apoptag Peroxidase In situ kit (Chemicon International, Temecula, CA) according to the manufacturer’s instructions. The entire area of each tumor core was evaluated. Apoptosis was quantified as the percentage of positively stained area over the whole tumor area.
Proliferation.
The expression of Ki-67, a cell proliferation marker, was assessed on 5-micron sections of the tumor tissue microarrays using a rabbit polyclonal antiboday (Vector Labs, Burlingame, CA; diluted 1:1500). After immunohistochemical staining, the tumor areas of each core were selected for quantitative image analysis using the Ariol instrument SL-50 (Applied Imaging, San Jose, CA). Cell proliferation was quantified as the percentage of Ki-67–positive nuclei over all tumor nuclei.
Statistical analysis
For the analyses of total incident prostate cancer, we calculated each individual’s person-time from date of return of the baseline questionnaire to the date of prostate cancer diagnosis, date of death from any cause, or the end of follow-up (January 31, 2010), whichever came first. For the analyses of lethal prostate cancer, we calculated each individual’s person-time from date of return of the baseline questionnaire to the date of metastasis or date of death or the end of follow-up. We excluded participants with history of cancer or with missing data on dietary intake of lycopene at baseline in 1986. For men who were diagnosed with prostate cancer, we only considered their diet before cancer diagnosis. Because approximately 90% of the HPFS participants are white, we did not perform stratified analyses by ethnicity.
All variables for dietary intake were energy adjusted using the residual method (38). To reduce intra-individual variation and represent long-term intake, we calculated cumulative updated average of lycopene intake as the primary measure of exposure. For example, the lycopene intake reported by the participants in 1986 was used to compute exposure in the 1986 to 1990 follow-up period, the average of lycopene intakes reported in 1986 and 1990 was used to compute exposure in the 1990 to 1994 follow-up period, and so on. We also used baseline lycopene intake (1986) and simply updated lycopene intake (most recent intake) as alternative measures of exposure. Variables of lycopene intakes were categorized into quintiles based on the total study population.
We used Cox proportional hazards regression models to estimate the hazard ratios (HRs) and 95% confidence intervals (CIs) for total and lethal prostate cancer. Models were stratified by year of questionnaire return and age at baseline (in months) and were adjusted for potential confounders that have been previously identified as risk factors of prostate cancer in this cohort. Interaction terms for lycopene intakes and age and year of questionnaire return were added into the models to test for the assumption of proportional hazards. Covariates included body mass index (BMI), height, family history of prostate cancer, race/ethnicity, smoking, vigorous physical activity, total energy, alpha-linolenic acid, calcium, and coffee. BMI, smoking, and vigorous physical activity were updated in each questionnaire circle, whereas cumulative updated average of intakes were used for total energy, alpha-linolenic acid, calcium, and coffee. To test for a linear trend across categories of intake, we modeled lycopene intake as a continuous variable using the median intake for each category.
To investigate whether PSA screening influenced the association between lycopene intake and prostate cancer incidence, we performed stratified survival analyses by time periods before and after the clinical introduction of PSA test in 1994.
To reduce the influence of PSA screening as a potential confounder of the relation between lycopene and total and lethal prostate cancer, we restricted in secondary analysis the study population to the participants who reported at least one negative PSA test by 2008. The follow-up for these participants started in 1994, when we first asked about PSA screening in the questionnaires, or in the year when their PSA testing was first reported. We then repeated the survival analyses to evaluate the associations between cumulative updated average of lycopene intake and subgroups of prostate cancer. To investigate whether the timing of diet had an effect on the association, we repeated the analyses using baseline or simply updated lycopene intake. We also put any two measures (for example, baseline and simply updated intake) simultaneously in the Cox regression models to explore the potentially different associations according to timing of the lycopene intake.
To gain insight into potential mechanisms as to how lycopene may affect prostate cancer progression, we evaluated levels of tissue biomarkers in the tumor specimen across the quintile of lycopene intake. These markers include characteristics of angiogenesis such as microvessel size and irregularity, apoptosis, and cell proliferation and differentiation. In addition to individual angiogenic markers, we also generated an angiogenic score based on microvessel density, diameter, area, and irregularity using factor analysis. This score has a mean of zero, and lower score reflects a greater angiogenic potential. We calculated a P value from analysis of variance for these markers across lycopene intake quintiles. To minimize the influence of tumor stage, we repeated the evaluation of angiogenic markers in tumors that were organ confined or with limited extraprostatic extension at prostatectomy (pT1–pT3a, N0, M0). We previously showed that among the markers of angiogenesis, vessel irregularity was the angiogenic feature most strongly predicting lethal prostate cancer (34), and therefore the association with this marker was our primary hypothesis.
All statistical analyses were two-sided and a P value of less than .05 was considered statistically significant. We conducted all analyses using the SAS software version 9.1 (SAS Institute, Cary, NC).
Results
Baseline height, BMI, family history of prostate cancer, smoking status, dietary intakes of total energy, carbohydrates, protein, alpha-linolenic acid, and calcium, and percentage of PSA screening by the year 2008, did not vary remarkably across quintiles of baseline lycopene intake (Table 1). Participants in the upper quintiles of lycopene intake were slightly younger and more likely to engage in vigorous physical activity. Men who consumed more lycopene in their diet also consumed less alcohol, coffee, and all three types of fats and slightly more fruits, vegetables, and dietary fiber. Dietary intake of lycopene was positively correlated with consumption of tomato and tomato products, such as tomato juice, tomato sauce, and pizza.
Table 1.
Lifestyle characteristics (at baseline unless otherwise indicated) according to the quintiles (Q) of dietary lycopene intake in the Health Professionals Follow-up Study cohort*
| Characteristics | Dietary lycopene intake, energy-adjusted (μg/day) | ||||
|---|---|---|---|---|---|
| Q1 | Q2 | Q3 | Q4 | Q5 | |
| 0–3687 | 3688–5301 | 5302–7062 | 7063–10130 | 10131–115012 | |
| No. of men | 9470 | 9609 | 9595 | 9598 | 9626 |
| Age, y | 56.8 (9.8) | 54.2 (9.7) | 53.2 (9.5) | 52.8 (9.6) | 52.9 (9.6) |
| Height, in | 70.2 (2.6) | 70.2 (2.6) | 70.2 (2.7) | 70.1 (2.7) | 70.0 (2.7) |
| BMI, kg/m2 | 25.4 (3.5) | 25.4 (3.2) | 25.5 (3.3) | 25.6 (3.3) | 25.8 (3.6) |
| White, % | 89.7 | 91.0 | 91.1 | 91.4 | 90.8 |
| Family history of prostate cancer, % | 11.7 | 11.3 | 11.0 | 10.7 | 10.0 |
| Smoking status, % | |||||
| Never smoker | 43.8 | 44.8 | 44.1 | 46.1 | 44.2 |
| Past smoker‚ quit ≤10 y | 13.3 | 12.6 | 12.8 | 12.7 | 12.8 |
| Past smoker‚ quit >10 y | 27.4 | 28.7 | 29.6 | 28.8 | 30.4 |
| Current smoker | 11.9 | 9.9 | 9.5 | 8.4 | 8.6 |
| Missing | 3.6 | 4.0 | 4.0 | 4.1 | 4.0 |
| Vigorous physical activity ≥10.5 METs/week, % | 28.7 | 31.0 | 34.3 | 36.2 | 37.7 |
| Ever PSA screening, % | 87.5 | 88.8 | 88.6 | 89.3 | 88.2 |
| Dietary intake of | |||||
| Total calories, kcal/day | 2058 (673) | 2071 (597) | 1951 (566) | 1903 (632) | 1949 (615) |
| Carbohydrates, g/day | 235 (91) | 240 (83) | 229 (79) | 226 (86) | 236 (87) |
| Protein, g/day | 93.4 (33.8) | 94.0 (29.4) | 89.9 (28.0) | 88.6 (30.3) | 91.7 (31.1) |
| Dietary fiber, g/day | 20.5 (9.2) | 21.9 (8.6) | 21.6 (8.6) | 22.0 (9.1) | 23.7 (10.0) |
| Total saturated fat, g/day | 26.6 (12.2) | 26.1 (10.5) | 23.9 (9.4) | 22.8 (10.1) | 22.3 (9.8) |
| Total monounsaturated fat, g/day | 29.4 (13.1) | 29.0 (11.2) | 26.6 (10.1) | 25.4 (10.8) | 25.3 (10.6) |
| Total polyunsaturated fat, g/day | 13.6 (6.5) | 13.7 (5.6) | 12.8 (5.1) | 12.6 (5.3) | 12.8 (5.1) |
| Total fruits, serving/day | 2.1 (1.5) | 2.3 (1.5) | 2.4 (1.5) | 2.5 (1.6) | 2.7 (1.8) |
| Total vegetables, serving/day | 2.3 (1.4) | 2.8 (1.4) | 3.0 (1.5) | 3.3 (1.7) | 4.0 (2.0) |
| Tomatoes, serving/week | 1.4 (1.4) | 2.2 (1.8) | 2.7 (2.0) | 3.2 (2.3) | 3.9 (2.9) |
| Tomato juice, serving/week | 0.1 (0.3) | 0.2 (0.4) | 0.3 (0.4) | 0.6 (0.8) | 1.5 (2.3) |
| Tomato sauce, serving/week | 0.3 (0.4) | 0.6 (0.3) | 0.7 (0.4) | 1.0 (0.8) | 2.2 (1.7) |
| Pizza, serving (2 slices)/week | 0.3 (0.3) | 0.5 (0.4) | 0.6 (0.5) | 0.7 (0.7) | 0.9 (1.0) |
| Alcohol, g/day | 12.4 (17.8) | 11.9 (15.9) | 11.2 (14.7) | 10.4 (14.0) | 10.7 (14.5) |
| α-Linolenic acid, g/day | 1.0 (0.4) | 1.1 (0.4) | 1.1 (0.4) | 1.1 (0.3) | 1.1 (0.3) |
| Calcium, mg/day | 910 (462) | 900 (417) | 894 (409) | 894 (420) | 889 (420) |
| Coffee, cup/day | 1.5 (1.8) | 1.4 (1.7) | 1.3 (1.6) | 1.2 (1.5) | 1.2 (1.5) |
* Values are means (SD) or percentages and are standardized to the age distribution of the study population. BMI = body mass index; PSA = prostate-specific antigen.
We found statistically significant inverse associations between quintiles of lycopene intake and incidence of total prostate cancer (P trend = .009), as well as lethal prostate cancer (P trend = .04) (Table 2). Compared with the bottom quintile, the top quintile of lycopene intake was associated with a hazard ratio of 0.91 (95% CI = 0.84 to 1.00) for total prostate cancer and a hazard ratio of 0.72 (95% CI = 0.56 to 0.94) for lethal prostate cancer. In the time period before the widespread use of PSA screening (before 1994), there was a borderline significant inverse association between lycopene intake and total prostate cancer (top quintile vs bottom quintile: HR = 0.85; 95% CI = 0.72 to 1.00; P trend = .07). However, during that period, men in the top quintile were at a 27.7% lower risk for lethal prostate cancer (HR = 0.72; 95% CI = 0.51 to 1.00; P trend = .07) compared with those in the bottom quintile of lycopene intake. In the PSA era, comparing participants in the top quintile with those in the bottom quintile of lycopene intake, there was a slight reduction in incidence of total prostate cancer, but this was less than that in the pre-PSA era. Only the lower risk for lethal prostate cancer remained unchanged (top quintile vs bottom quintile: HR = 0.72; 95% CI = 0.49 to 1.10; P trend = .10).
Table 2.
Hazard ratios (HRs) and 95% confidence interval (CIs) of total and lethal prostate cancer in relation to lycopene intake in the Health Professionals Follow-up Study (1986–January, 2010) cohort
| Prostate cancer and time period | Cumulative updated average of dietary lycopene intake, HR (95% CI) | P trend* | ||||
|---|---|---|---|---|---|---|
| Q1 | Q2 | Q3 | Q4 | Q5 | ||
| Total prostate cancer | 1209 | 1236 | 1144 | 1102 | 1037 | |
| Adjusted for age | 1.00 (referent) | 1.10 (1.00 to 1.20) | 1.00 (0.95 to 1.10) | 1.00 (0.94 to 1.10) | 0.96 (0.88 to 1.00) | .07 |
| Multivariable adjusted† | 1.00 (referent) | 1.00 (0.96 to 1.10) | 0.97 (0.90 to 1.10) | 0.97 (0.89 to 1.10) | 0.93 (0.86 to 1.00) | .03 |
| Multivariable adjusted‡ | 1.00 (referent) | 1.00 (0.95 to 1.10) | 0.96 (0.89 to 1.00) | 0.96 (0.88 to 1.00) | 0.91 (0.84 to 1.00) | .009 |
| Lethal prostate cancer | 184 | 136 | 118 | 128 | 92 | |
| Adjusted for age | 1.00 (referent) | 0.92 (0.74 to 1.10) | 0.86 (0.68 to 1.10) | 0.98 (0.78 to 1.20) | 0.72 (0.56 to 0.92) | .02 |
| Multivariable adjusted† | 1.00 (referent) | 0.91 (0.73 to 1.10) | 0.85 (0.68 to 1.10) | 1.00 (0.81 to 1.30) | 0.75 (0.58 to 0.97) | .08 |
| Multivariable adjusted‡ | 1.00 (referent) | 0.90 (0.72 to 1.10) | 0.84 (0.66 to 1.10) | 0.99 (0.78 to 1.20) | 0.72 (0.56 to 0.94) | .04 |
| Pre–prostate-specific antigen era | ||||||
| Total prostate cancer | 420 | 310 | 280 | 262 | 241 | |
| Adjusted for age | 1.00 (referent) | 0.92 (0.79 to 1.10) | 0.92 (0.79 to 1.10) | 0.93 (0.80 to 1.10) | 0.85 (0.72 to 0.99) | .06 |
| Multivariable adjusted† | 1.00 (referent) | 0.93 (0.80 to 1.10) | 0.91 (0.78 to 1.10) | 0.95 (0.81 to 1.10) | 0.87 (0.74 to 1.00) | .10 |
| Multivariable adjusted‡ | 1.00 (referent) | 0.92 (0.79 to 1.10) | 0.90 (0.77 to 1.00) | 0.93 (0.80 to 1.10) | 0.85 (0.72 to 1.00) | .07 |
| Lethal prostate cancer | 104 | 73 | 63 | 72 | 48 | |
| Adjusted for age | 1.00 (referent) | 0.93 (0.69 to 1.30) | 0.86 (0.63 to 1.20) | 1.10 (0.79 to 1.4) | 0.71 (0.50 to 1.00) | .10 |
| Multivariable adjusted† | 1.00 (referent) | 0.93 (0.69 to 1.30) | 0.89 (0.65 to 1.20) | 1.10 (0.84 to 1.5) | 0.76 (0.54 to 1.10) | .30 |
| Multivariable adjusted‡ | 1.00 (referent) | 0.93 (0.68 to 1.30) | 0.88 (0.64 to 1.20) | 1.10 (0.81 to 1.5) | 0.72 (0.51 to 1.00) | .20 |
| Prostate-specific antigen era | ||||||
| Total prostate cancer | 789 | 926 | 864 | 840 | 796 | |
| Adjusted for age | 1.00 (referent) | 1.20 (1.10 to 1.30) | 1.10 (0.99 to 1.20) | 1.10 (0.97 to 1.20) | 1.00 (0.91 to 1.10) | .30 |
| Multivariable adjusted† | 1.00 (referent) | 1.10 (0.99 to 1.20) | 1.00 (0.91 to 1.10) | 0.99 (0.89 to 1.10) | 0.96 (0.87 to 1.10) | .09 |
| Multivariable adjusted‡ | 1.00 (referent) | 1.10 (0.98 to 1.20) | 1.00 (0.90 to 1.10) | 0.98 (0.89 to 1.10) | 0.95 (0.86 to 1.10) | .05 |
| Lethal prostate cancer | 80 | 63 | 55 | 56 | 44 | |
| Adjusted for age | 1.00 (referent) | 0.90 (0.65 to 1.30) | 0.85 (0.60 to 1.20) | 0.89 (0.63 to 1.30) | 0.72 (0.50 to 1.10) | .10 |
| Multivariable adjusted† | 1.00 (referent) | 0.88 (0.63 to 1.20) | 0.81 (0.57 to 1.10) | 0.88 (0.62 to 1.20) | 0.74 (0.50 to 1.10) | .10 |
| Multivariable adjusted‡ | 1.00 (referent) | 0.87 (0.62 to 1.20) | 0.79 (0.56 to 1.10) | 0.86 (0.61 to 1.20) | 0.72 (0.49 to 1.10) | .10 |
* P values were calculated from two-sided χ2 tests by modeling lycopene intake as a continuous variable using the median intake for each category.
† Hazard ratios adjusted for age (months), height (<66, 66.0–67.9, 68.0–69.9, 70.0–71.9, ≥72.0 inches), body mass index (<21.0, 21.0–22.9, 23.0–24.9, 25.0–27.4, 27.5–29.9, ≥30.0kg/m2), race (black, white, Asian, other), family history of prostate cancer (yes or no), vigorous activity (none, 0–3.4, 3.5–10.4, 10.5–28.4, ≥28.5 MET hours/week), smoking status (never; past and quit >10 years; past and quit ≤10 years; current; missing), and caloric intake (kcal/day, quintiles).
‡ Hazard ratios adjusted for age (months), height (<66, 66.0–67.9, 68.0–69.9, 70.0–71.9, ≥70.0 inches), body mass index (<21.0, 21.0–22.9, 23.0–24.9, 25–27.4, 27.5–29.9, ≥30.0kg/m2), race (black, white, Asian, other), family history of prostate cancer (yes or no), vigorous activity (none, 0–3.4, 3.5–10.4, 10.5–28.4, ≥28.5 MET hours/week), smoking status (never; past and quit >10 years; past and quit ≤10 years; current; missing), and dietary intakes of calcium (500–749, 750–999, 1000–1499, 1500–1999, ≥2000mg/day), α-linolenic acid (g/day, quintiles), coffee (none, <1, 1–3, 4–5, ≥6 cups/day), and total calories (kcal/day, quintiles).
To further reduce the influence of PSA screening on the observed association, in a subanalysis we restricted the study population to men who had at least one negative PSA test. In addition to cumulative updated average of lycopene intake, we also evaluated baseline and simply updated lycopene intake as alternative measures of exposure. Because the results from the age-adjusted analyses were similar to those from full multivariable models, we only present the multivariable hazard ratios (Table 3). Results from the multivariable models adjusting for various dietary factors were comparable with those from the models without adjusting for them. The strongest associations were observed for lethal prostate cancer with baseline or cumulative updated average of lycopene intake (top quintile vs bottom quintile: HR = 0.48, 95% CI = 0.30 to 0.78, P trend = .009 for baseline lycopene intake; HR = 0.47, 95% CI = 0.29 to 0.75, P trend = .009 for cumulative updated lycopene intake). Men with the highest intake were half as likely to develop lethal prostate cancer compared with those with the lowest intake. The association was much weaker for the simply updated diet, reflecting most recent intake, suggesting that lycopene may be acting early in the disease process.
Table 3.
Hazard ratios (HR) and 95% confidence intervals (CIs) of total and lethal prostate cancer in relation to different measurements of dietary lycopene intake in the Health Professionals Follow-up Studyy (1994–January, 2010), among men who had at least one negative prostate-specific antigen test
| Measures of dietary lycopene intake | Dietary lycopene intake, HR (95% CI) | P trend* | ||||
|---|---|---|---|---|---|---|
| Q1 | Q2 | Q3 | Q4 | Q5 | ||
| Total prostate cancer | ||||||
| Cumulative average, updated† | ||||||
| No. of cases | 779 | 803 | 748 | 747 | 691 | |
| Multivariable-adjusted‡ | 1.00 (referent) | 0.98 (0.89 to 1.10) | 0.91 (0.82 to 1.00) | 0.93 (0.84 to 1.00) | 0.90 (0.81 to 1.00) | .03 |
| Multivariable-adjusted§ | 1.00 (referent) | 0.98 (0.89 to 1.10) | 0.90 (0.82 to 1.00) | 0.92 (0.83 to 1.00) | 0.88 (0.79 to 0.98) | .02 |
| Simply updated║ | ||||||
| No. of cases | 816 | 791 | 757 | 705 | 699 | |
| Multivariable-adjusted‡ | 1.00 (referent) | 0.97 (0.88 to 1.10) | 0.99 (0.89 to 1.10) | 0.95 (0.86 to 1.10) | 0.93 (0.84 to 1.00) | .10 |
| Multivariable-adjusted§ | 1.00 (referent) | 0.97 (0.88 to 1.10) | 0.98 (0.89 to 1.10) | 0.95 (0.85 to 1.00) | 0.92 (0.83 to 1.00) | .09 |
| Baseline¶ | ||||||
| No. of cases | 773 | 772 | 730 | 663 | ||
| Multivariable-adjusted‡ | 1.00 (referent) | 0.94 (0.85 to 1.00) | 0.98 (0.88 to 1.10) | 0.93 (0.84 to 1.00) | 0.87 (0.78 to 0.96) | .009 |
| Multivariable-adjusted§ | 1.00 (referent) | 0.93 (0.85 to 1.00) | 0.97 (0.88 to 1.10) | 0.92 (0.83 to 1.00) | 0.86 (0.77 to 0.95) | .004 |
| Lethal prostate cancer | ||||||
| Cumulative average, updated§ | ||||||
| No. of cases | 82 | 59 | 51 | 63 | 31 | |
| Multivariable-adjusted‡ | 1.00 (referent) | 0.75 (0.51 to 1.10) | 0.65 (0.43 to 0.98) | 0.90 (0.61 to 1.30) | 0.48 (0.29 to 0.77) | .01 |
| Multivariable-adjusted§ | 1.00 (referent) | 0.74 (0.50 to 1.10) | 0.64 (0.43 to 0.96) | 0.88 (0.60 to 1.30) | 0.47 (0.29 to 0.75) | .009 |
| Simply updated║ | ||||||
| No. of cases | 73 | 72 | 49 | 53 | 39 | |
| Multivariable-adjusted‡ | 1.00 (referent) | 1.10 (0.73 to 1.60) | 0.85 (0.56 to 1.30) | 0.95 (0.63 to 1.40) | 0.67 (0.42 to 1.10) | .07 |
| Multivariable-adjusted§ | 1.00 (referent) | 1.10 (0.73 to 1.50) | 0.84 (0.56 to 1.30) | 0.94 (0.62 to 1.40) | 0.66 (0.42 to 1.10) | .06 |
| Baseline¶ | ||||||
| No. of cases | 83 | 62 | 51 | 58 | 32 | |
| Multivariable-adjusted‡ | 1.00 (referent) | 0.82 (0.56 to 1.20) | 0.74 (0.50 to 1.10) | 0.93 (0.63 to 1.40) | 0.50 (0.31 to 0.80) | .01 |
| Multivariable-adjusted§ | 1.00 (referent) | 0.81 (0.55 to 1.20) | 0.74 (0.49 to 1.10) | 0.92 (0.62 to 1.40) | 0.48 (0.30 to 0.78) | .009 |
* P values were calculated from two-sided χ2 tests by modeling lycopene intake as a continuous variable using the median intake for each category.
† Cumulative average of lycopene intake from 1986 to 2006 or to the questionnaire year before cancer diagnosis.
‡ Hazard ratios adjusted for age (months), height (<66.0, 66.0–67.9, 68.0–69.9, 70.0–71.9, ≥70 inches), body mass index (<21.0, 21.0–22.9, 23.0–24.9, 25.0–27.4, 27.5–29.9, ≥30.0 kg/m2), race (black, white, Asian, other), family history of prostate cancer (yes or no), vigorous activity (none, 0–3.4, 3.5–10.4, 10.5–28.4, ≥28.5 MET hours/week), smoking status (never; past and quit >10 years; past and quit ≤10 years; current; missing), and caloric intake (kcal/day, quintiles).
§ Hazard ratios adjusted for age (months), height (<66.0, 66.0–67.9, 68.0–69.9, 70.0–71.9, ≥72.0 inches), body mass index (<21.0, 21.0–22.9, 23.0–24.9, 25.0–27.4, 27.5–29.9, ≥30 kg/m2), race (black, white, Asian, other), family history of prostate cancer (yes or no), vigorous activity (none, 0–3.4, 3.5–10.4, 10.5–28.4, ≥28.5 MET hours/week), smoking status (never; past and quit >10 years; past and quit ≤10 years; current; missing), and dietary intakes of calcium (500–749, 750–999, 1000–1499, 1500–1999, ≥2000 mg/day), α-linolenic acid (g/day, quintiles), coffee (none, <1, 1–3, 4–5, ≥6 cups/day), and total calories (kcal/day, quintiles).
║ The most recent lycopene intake before cancer diagnosis.
¶ Lycopene intake in 1986.
To evaluate which measure of exposure best captures the protective association with lycopene, we put any two measures simultaneously in the Cox regression models (Table 4). The multivariable models with or without dietary intakes of calcium, alpha-linolenic acid, and coffee yielded similar results. The associations persisted for baseline intake and cumulative average of intake and did not persist for simply updated intake. These results support that measures of long-term or early dietary lycopene intake are more relevant for prevention against prostate cancer, especially aggressive cancers.
Table 4.
Hazard ratios (HRs) and 95% confidence intervals (CIs) of prostate cancer in relation to different measurements of dietary lycopene intake in the Health Professionals Follow-up Study (1994–January, 2010) among men who had at least one negative prostate-specific antigen test
| Models | Dietary lycopene intake, HR (95% CI) | P trend* | ||||
|---|---|---|---|---|---|---|
| Q1 | Q2 | Q3 | Q4 | Q5 | ||
| Total prostate cancer | ||||||
| Model 1† | ||||||
| Cumulative average, updated‡ | 1.00 (referent) | 0.97 (0.87 to 1.10) | 0.89 (0.79 to 1.00) | 0.92 (0.80 to 1.00) | 0.89 (0.76 to 1.00) | .10 |
| Simply updated§ | 1.00 (referent) | 1.00 (0.90 to 1.10) | 1.00 (0.93 to 1.20) | 1.00 (0.90 to 1.20) | 1.00 (0.87 to 1.20) | 1.00 |
| Model 2║ | ||||||
| Cumulative average, updated‡ | 1.00 (referent) | 0.97 (0.87 to 1.10) | 0.89 (0.78 to 1.00) | 0.90 (0.79 to 1.00) | 0.87 (0.75 to 1.00) | .08 |
| Simply updated§ | 1.00 (referent) | 0.99 (0.90 to 1.10) | 1.00 (0.93 to 1.20) | 1.00 (0.90 to 1.20) | 1.00 (0.88 to 1.20) | .90 |
| Model 3† | ||||||
| Baseline¶ | 1.00 (referent) | 0.94 (0.85 to 1.00) | 0.98 (0.88 to 1.10) | 0.94 (0.84 to 1.00) | 0.88 (0.78 to 0.98) | .03 |
| Simply updated§ | 1.00 (referent) | 0.99 (0.89 to 1.10) | 1.00 (0.90 to 1.10) | 0.98 (0.88 to 1.10) | 0.97 (0.87 to 1.10) | .60 |
| Model 4║ | ||||||
| Baseline¶ | 1.00 (referent) | 0.94 (0.85 to 1.00) | 0.97 (0.88 to 1.10) | 0.93 (0.83 to 1.00) | 0.87 (0.78 to 0.97) | .02 |
| Simply updated§ | 1.00 (referent) | 0.99 (0.89 to 1.10) | 1.00 (0.90 to 1.10) | 0.98 (0.88 to 1.10) | 0.97 (0.86 to 1.10) | .50 |
| Model 5† | ||||||
| Cumulative average, updated‡ | 1.00 (referent) | 1.00 (0.89 to 1.10) | 0.92 (0.82 to 1.00) | 0.97 (0.85 to 1.10) | 0.96 (0.83 to 1.10) | .70 |
| Baseline¶ | 1.00 (referent) | 0.95 (0.85 to 1.10) | 1.00 (0.90 to 1.10) | 0.96 (0.85 to 1.10) | 0.89 (0.78 to 1.00) | .10 |
| Model 6║ | ||||||
| Cumulative average, updated‡ | 1.00 (referent) | 0.99 (0.89 to 1.10) | 0.92 (0.81 to 1.00) | 0.96 (0.84 to 1.10) | 0.95 (0.82 to 1.10) | .50 |
| Baseline¶ | 1.00 (referent) | 0.95 (0.85 to 1.10) | 1.00 (0.89 to 1.10) | 0.96 (0.84 to 1.10) | 0.89 (0.77 to 1.00) | .10 |
| Lethal prostate cancer | ||||||
| Model 1† | ||||||
| Cumulative average, updated‡ | 1.00 (referent) | 0.66 (0.42 to 1.00) | 0.58 (0.35 to 0.97) | 0.80 (0.46 to 1.40) | 0.45 (0.22 to 0.90) | .07 |
| Simply updated§ | 1.00 (referent) | 1.30 (0.86 to 2.00) | 1.20 (0.70 to 2.00) | 1.30 (0.71 to 2.20) | 1.10 (0.56 to 2.20) | .80 |
| Model 2║ | ||||||
| Cumulative average, updated‡ | 1.00 (referent) | 0.65 (0.42 to 1.00) | 0.57 (0.34 to 0.95) | 0.79 (0.45 to 1.40) | 0.44 (0.22 to 0.88) | .06 |
| Simply updated§ | 1.00 (referent) | 1.30 (0.86 to 2.00) | 1.20 (0.71 to 2.00) | 1.30 (0.71 to 2.20) | 1.10 (0.56 to 2.20) | .80 |
| Model 3† | ||||||
| Baseline¶ | 1.00 (referent) | 0.80 (0.54 to 1.20) | 0.76 (0.49 to 1.20) | 0.94 (0.62 to 1.40) | 0.53 (0.32 to 0.90) | .05 |
| Simply updated§ | 1.00 (referent) | 1.20 (0.78 to 1.70) | 0.96 (0.62 to 1.50) | 1.10 (0.67 to 1.70) | 0.85 (0.51 to 1.40) | .50 |
| Model 4║ | ||||||
| Baseline¶ | 1.00 (referent) | 0.79 (0.53 to 1.20) | 0.75 (0.49 to 1.20) | 0.93 (0.61 to 1.40) | 0.52 (0.31 to 0.88) | .04 |
| Simply updated§ | 1.00 (referent) | 1.20 (0.78 to 1.70) | 0.95 (0.61 to 1.50) | 1.00 (0.67 to 1.60) | 0.85 (0.51 to 1.40) | .40 |
| Model 5† | ||||||
| Cumulative average, updated‡ | 1.00 (referent) | 0.77 (0.50 to 1.20) | 0.68 (0.41 to 1.10) | 0.94 (0.55 to 1.60) | 0.62 (0.31 to 1.20) | .40 |
| Baseline¶ | 1.00 (referent) | 0.94 (0.61 to 1.20) | 0.91 (0.55 to 1.50) | 1.10 (0.63 to 1.80) | 0.66 (0.34 to 1.30) | .30 |
| Model 6║ | ||||||
| Cumulative average, updated‡ | 1.00 (referent) | 0.77 (0.49 to 1.20) | 0.67 (0.40 to 1.10) | 0.93 (0.55 to 1.60) | 0.61 (0.31 to 1.20) | .30 |
| Baseline¶ | 1.00 (referent) | 0.92 (0.60 to 1.40) | 0.90 (0.55 to 1.50) | 1.10 (0.62 to 1.80) | 0.65 (0.33 to 1.30) | .30 |
* P values were calculated from two-sided χ2 tests by modeling lycopene intake as a continuous variable using the median intake for each category.
† Hazard ratios adjusted for age (months), height (<66.0, 66.0–67.9, 68.0–69.9, 70.0–71.9, ≥70 inches), body mass index (<21.0, 21.0–22.9, 23.0–24.9, 25.0–27.4, 27.5–29.9, ≥30.0kg/m2), race (black, white, Asian, other), family history of prostate cancer (yes or no), vigorous activity (none, 0–3.4, 3.5–10.4, 10.5–28.4, ≥28.5 MET hours/week), smoking status (never; past and quit >10 years; past and quit ≤10 years; current; missing), and caloric intake (kcal/day, quintiles).
‡ Cumulative average of lycopene intake from 1986 to 2006 or to the questionnaire year before cancer diagnosis.
§ The most recent lycopene intake before cancer diagnosis.
║ Hazard ratios adjusted for age (months), height (<66.0, 66.0–67.9, 68.0–69.9, 70.0–71.9, ≥72.0 inches), body mass index (<21.0, 21.0–22.9, 23.0–24.9, 25.0–27.4, 27.5–29.9, ≥30 kg/m2), race (black, white, Asian, other), family history of prostate cancer (yes or no), vigorous activity (none, 0–3.4, 3.5–10.4, 10.5–28.4, ≥28.5 MET hours/week), smoking status (never; past and quit >10 years; past and quit ≤10 years; current; missing), and dietary intakes of calcium (500–749, 750–999, 1000–1499, 1500–1999, ≥2000 mg/day), α-linolenic acid (g/day, quintiles), coffee (none, <1, 1–3, 4–5, ≥6 cups/day), and total calories (kcal/day, quintiles).
¶ Lycopene intake in 1986.
Table 5 presents the association between quintiles of lycopene intake and levels of tumor markers for angiogenesis, apoptosis, and cellular proliferation and differentiation. There was no association between intake and extent of tumor apoptosis and cellular proliferation or differentiation. However, three of the tumor angiogenesis markers, microvessel diameter and area and irregularity of the vessel lumen, as well as the summary angiogenic score, were strongly associated with lycopene intake in such a way that men with higher lycopene intake had tumors that displayed less angiogenic potential. Because lycopene intake and angiogenesis are both associated with tumor stage, we evaluated the relation between angiogenic measures and lycopene intake in tumors confined within the prostate or with limited extraprostatic extension (T1–T3a, N0, M0) to minimize the effect of tumor stage. Statistically significant trends persisted for all three angiogenic markers and the score.
Table 5.
Tumor markers (angiogenesis, apoptosis, proliferation, and differentiation) by quintiles of dietary lycopene intake in the Health Professionals Follow-up Study cohort
| Tumor biomarkers | Overall | Dietary Lycopene intake (cumulative average, updated) | P trend* | ||||
|---|---|---|---|---|---|---|---|
| Q1 | Q2 | Q3 | Q4 | Q5 | |||
| All tumors | |||||||
| Angiogenesis–CD34 staining | 570 | 136 | 116 | 118 | 114 | 86 | |
| Vessel diameter†, µm, mean (SD) | 12.0 (2.5) | 11.2 (2.2) | 12.1 (2.4) | 12.2 (2.6) | 12.3 (3.0) | 12.3 (2.0) | .0007 |
| Vessel area†, µm2, mean (SD) | 121.4 (62.4) | 101.7 (48.3) | 124.5 (61.9) | 125.0 (59.7) | 133.0 (81.3) | 128.4 (50.7) | .0001 |
| Irregularity of vessel lumen‡, mean (SD) | 4.03 (1.1) | 4.3 (1.0) | 4.2 (1.1) | 3.8 (1.0) | 3.9 (1.1) | 3.9 (1.2) | .0002 |
| Angiogenic score§, mean (SD) | 0 (0.95) | −0.32 (0.78) | 0.04 (0.91) | 0.07 (0.94) | 0.16 (1.21) | 0.12 (0.77) | .0007 |
| Apoptosis–TUNEL assay | 454 | 106 | 98 | 97 | 93 | 60 | |
| Percentage of stained area, mean (SD) | 2.2 (4.7) | 2.6 (5.3) | 1.9 (3.9) | 2.0 (4.1) | 1.5 (3.4) | 3.2 (1.0) | .70 |
| Cell proliferation–Ki-67 staining | 372 | 85 | 70 | 83 | 79 | 55 | |
| Number of positive nuclei, mean (SD) | 10.5 (32.4) | 7.9 (11.7) | 13.1 (40.1) | 10.3 (23.4) | 7.1 (8.2) | 16.6 (62.7) | .80 |
| Percentage of positive nuclei, mean (SD) | 0.75 (1.7) | 0.61 (0.98) | 0.73 (1.4) | 0.89 (2.0) | 0.58 (0.89) | 1.0 (2.8) | .90 |
| Cell differentiation–tumor grade | 1007 | 207 | 202 | 216 | 209 | 173 | |
| Gleason score, mean (SD) | 7.2 (0.93) | 7.2 (0.91) | 7.2 (0.95) | 7.3 (0.97) | 7.2 (0.93) | 7.1 (0.88) | .60 |
| Organ-confined tumors | |||||||
| Angiogenesis–CD34 staining | 360 | 82 | 73 | 75 | 71 | 59 | |
| Vessel diameter†, µm, mean (SD) | 12.2 (2.5) | 11.3 (2.0) | 12.2 (2.3) | 12.5 (2.8) | 12.9 (3.0) | 12.1 (1.8) | .01 |
| Vessel area†, µm2, mean (SD) | 125.7 (64.6) | 102.1 (42.6) | 126.1 (65.5) | 132.7 (63.7) | 147.4 (87.5) | 122.8 (46.1) | .002 |
| Irregularity of vessel lumen‡, mean (SD) | 4.0 (1.1) | 4.3 (1.0) | 4.2 (1.1) | 3.7 (1.0) | 3.7 (1.1) | 3.8 (1.1) | .0001 |
| Angiogenic score§, mean (SD) | 0.07 (0.96) | −0.29 (0.70) | 0.07 (0.93) | 0.19 (1.0) | 0.39 (1.3) | 0.03 (0.7) | .007 |
* P trend was calculated using analysis of variance and was two-sided.
† Smaller vessels are more angiogenic.
‡ Irregular vessels are more angiogenic; a score of 1.0 indicates a perfect circle.
§ Lower angiogenic scores are more angiogenic.
Discussion
In this study, lycopene intake was inversely associated with incidence of total prostate cancer and lethal prostate cancer. In addition, we evaluated tumor biomarkers for various cellular and molecular events in relation to lycopene intake and found that higher lycopene intake was associated with lower angiogenic potential in the tumor based on the vessel size and shape. Based on these results, we hypothesize that the consumption of a diet rich in lycopene-containing foods reduces the aggressive potential of prostate cancer by inhibiting the neoangiogenesis that occurs in tumor development.
Several factors may have contributed to the fact that we observed the strongest association for lycopene intake in the analysis restricted to those with a negative PSA test at baseline. In this analytic design, subsequent diagnosis of prostate cancer, typically resulting from a rise in PSA, may represent a true change in the activity of the cancer. In studies where there are many prevalent cases, as in the initial use of PSA screening, the cancer may have been prevalent for many years and the exposure over the study period would not be expected to influence the cancer. Further, the potential for detection bias is diminished in this analysis.
Because of mixed evidence in the literature, the role of tomato products and lycopene in prostate cancer etiology and prevention remained controversial. In 2007, the US Food and Drug Administration published an evidence-based review of tomato, lycopene, and cancer and concluded that there is “very limited evidence to support an association between tomato consumption and reduced risk of prostate cancer” (39). In contrast, several months later, the World Cancer Research Fund suggested that their systematic review of the literature supported a likely relationship (40). Findings in our study and others support an inverse association between lycopene-rich foods and prostate cancer, particularly for lethal disease. Several factors may contribute to this inconsistency.
First, measurement for bioavailable lycopene may be imprecise in many studies (41). The bioavailability of lycopene varies greatly in food sources. When tomato and certain fruits are consumed raw, the absorption of lycopene may be lower. Food processing and cooking greatly enhances the bioavailability of lycopene by disrupting its binding to matrices and making the highly lipophilic lycopene readily available for intestinal absorption (42,43). Our dietary assessments of lycopene intake, capturing a fourfold increase in the mean values across quintiles, corresponded to a substantial difference (37%) in plasma lycopene concentrations.
Second, the dietary intake of lycopene was repeatedly measured in our cohort, compared with a single assessment in the majority of other studies. With the use of cumulative average of lycopene intake, we were able to capture the changes in diet over time and to reduce within-person variation. We also compared different measures of dietary lycopene intake, including baseline and simply updated lycopene intake, in addition to the cumulative average. Lycopene intake at baseline yielded an inverse association with prostate cancer similar to that with the cumulative average, suggesting that long-term or remote lycopene intake was more etiologically relevant than recent intake.
Third, the distributions of lycopene intake vary greatly across populations. In our study, the median values of energy-adjusted lycopene intake for quintiles were 3160, 5101, 6744, 8923, and 13391 μg/day, with a fourfold difference in the median values between the highest and lowest quintiles. In a hospital-based case–control study in Uruguay (14), the investigators evaluated various foods and nutrients in relation to risk of prostate cancer and observed no association with lycopene. However, the ranges of lycopene intake quartiles were 1300 or less, 1301 to 2501, 2502 to 3300, and more than 3301 μg/day. Dietary intake of lycopene in this population may have been too low to be informative.
Fourth, the association between lycopene intake and prostate cancer risk was weak for endpoints enriched with indolent cancers (eg, total prostate cancer in the PSA era), stronger for lethal prostate cancer than for total prostate cancer, and particularly strong when restricted to men who have had at least one negative PSA test at baseline. This overall pattern suggests that lycopene may be primarily influencing progression of prostate cancers. Similar to our results, studies conducted in the United States before the widespread use of PSA screening (4) or in other areas where the PSA screening is not prevalent (18,24) generally support a positive role of lycopene in prostate cancer prevention. In settings where incident cases were primarily diagnosed through PSA screening, the studies generally yielded null findings (7,12). Further, in the observational study component of the Prostate Cancer Prevention Trial (8,19), a reanalysis showed no association between lycopene level and cancers that had shown no evidence of progression but were detected by an end-of-study biopsy, but lycopene was associated with cancers that showed signs of progression during the study through symptoms, growth, or rise in PSA (44).
Interestingly, corresponding to the suggestion that lycopene influences primarily progression of prostate cancer, we found that dietary lycopene intake was correlated with less angiogenic potential in the tumor. Previously in the HPFS, we reported that microvessel morphological markers, such as microvessel diameter and area, and the irregularity of the vessel lumen, strongly predict lethal prostate cancer (34). Of note, angiogenic factors correlated with stage and lethality largely independently of grade, suggesting that factors that influence progression through angiogenesis would do so largely independently of tumor grade.
To our knowledge, this is the first epidemiological study to report the potential antiangiogenic effect of lycopene-rich foods. Lycopene may inhibit angiogenesis of prostate cancer cells by regulating vascular endothelial growth factor (45). Elgass et al., using human umbilical vein endothelial cells, reported that lycopene inhibited in vitro angiogenesis at physiologically relevant concentrations (46). Huang et al. reported that the antiangiogenic effects of lycopene were mediated through immunomodulation of proangiogenic cytokine secretion in human peripheral blood mononuclear cells (47). Chen et al. showed the antiangiogenic activity of lycopene both in vitro and in vivo and proposed that the mechanism may involve PI3K-Akt and ERK/p38 signaling pathways (48).
As for all observational studies, the main limitation is the potential for uncontrolled confounding by unknown factors. Although this possibility could not be ruled out entirely, the likelihood that uncontrolled confounders entirely accounted for our results on lycopene and prostate cancer is low. Humans consume lycopene through specific food sources, such as tomato and tomato-based products, which is only a small component of the entire diet. These foods are not strongly associated with other lifestyle or dietary factors (see Table 1). Age-adjusted analyses, multivariable analyses adjusting for lifestyle factors, and multivariable analyses adjusting for both lifestyle and dietary factors yielded very similar results. Measurement error is also a factor in nutritional studies, but as discussed above, several features of our study design, including repeated measures, may have minimized error. Finally, our cohort is comprised of mainly white health professionals, which may limit generalizability of our results.
In conclusion, our results suggest the intake of lycopene is associated with reduced risk of lethal prostate cancer. In the setting of widespread PSA screening, advanced or lethal prostate cancer is preferable to total prostate cancer as the endpoint for epidemiological studies of lycopene.
Funding
This project was funded by grants from the National Institutes of Health (R01 CA133891 to EG; CA136578 to LM; CA141298 to MJS). The Health Professionals Follow-up Study was supported by the National Institutes of Health/National Cancer Institute (UM1 CA167552-01).
The study sponsors had no role in the design of the study; the collection, analysis, and interpretation of the data; the writing of the manuscript; and the decision to submit the manuscript for publication.
We would like to thank the participants and staff of the Health Professionals Follow-up Study, for their valuable contributions, as well as the following state cancer registries for their help: AL, AZ, AR, CA, CO, CT, DE, FL, GA, ID, IL, IN, IA, KY, LA, ME, MD, MA, MI, NE, NH, NJ, NY, NC, ND, OH, OK, OR, PA, RI, SC, TN, TX, VA, WA, WY.
References
- 1. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2012. CA Cancer J Clin. 2012;62(1):10–29 [DOI] [PubMed] [Google Scholar]
- 2. Giovannucci E. Does prostate-specific antigen screening influence the results of studies of tomatoes, lycopene, and prostate cancer risk? J Natl Cancer Inst. 2007;99(14):1060–1062 [DOI] [PubMed] [Google Scholar]
- 3. Giovannucci E, Liu Y, Platz EA, et al. Risk factors for prostate cancer incidence and progression in the health professionals follow-up study. Int J Cancer. 2007;121(7):1571–1578 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Ilic D, Forbes KM, Hassed C. Lycopene for the prevention of prostate cancer. Cochrane Database Syst Rev. 2011;2011(11):CD008007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Giovannucci E, Ascherio A, Rimm EB, et al. Intake of carotenoids and retinol in relation to risk of prostate cancer. J Natl Cancer Inst. 1995;87(23):1767–1776 [DOI] [PubMed] [Google Scholar]
- 6. Cerhan JR, Cantor KP, Williamson K, et al. Cancer mortality among Iowa farmers: recent results, time trends, and lifestyle factors (United States). Cancer Causes Control. 1998;9(3):311–319 [DOI] [PubMed] [Google Scholar]
- 7. Mills PK, Beeson WL, Phillips RL, et al. Cohort study of diet, lifestyle, and prostate cancer in Adventist men. Cancer. 1989;64(3):598–604 [DOI] [PubMed] [Google Scholar]
- 8. Kirsh VA, Mayne ST, Peters U, et al. A prospective study of lycopene and tomato product intake and risk of prostate cancer. Cancer Epidemiol Biomarkers Prev. 2006;15(1):92–98 [DOI] [PubMed] [Google Scholar]
- 9. Kristal AR, Arnold KB, Neuhouser ML, et al. Diet, supplement use, and prostate cancer risk: results from the prostate cancer prevention trial. Am J Epidemiol. 2010;172(5):566–577 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Giovannucci E, Rimm EB, Liu Y, et al. A prospective study of tomato products, lycopene, and prostate cancer risk. J Natl Cancer Inst. 2002;94(5):391–398 [DOI] [PubMed] [Google Scholar]
- 11. Le Marchand L, Hankin JH, Kolonel LN, et al. Vegetable and fruit consumption in relation to prostate cancer risk in Hawaii: a reevaluation of the effect of dietary beta-carotene. Am J Epidemiol. 1991;133(3):215–219 [DOI] [PubMed] [Google Scholar]
- 12. Key TJ, Silcocks PB, Davey GK, et al. A case-control study of diet and prostate cancer. Br J Cancer. 1997;76(5):678–687 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Meyer F, Bairati I, Fradet Y, et al. Dietary energy and nutrients in relation to preclinical prostate cancer. Nutr Cancer. 1997;29(2):120–126 [DOI] [PubMed] [Google Scholar]
- 14. Deneo-Pellegrini H, De Stefani E, Ronco A, et al. Foods, nutrients and prostate cancer: a case-control study in Uruguay. Br J Cancer. 1999;80(3–4):591–597 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Jain MG, Hislop GT, Howe GR, et al. Plant foods, antioxidants, and prostate cancer risk: findings from case-control studies in Canada. Nutr Cancer. 1999;34(2):173–184 [DOI] [PubMed] [Google Scholar]
- 16. Vogt TM, Mayne ST, Graubard BI, et al. Serum lycopene, other serum carotenoids, and risk of prostate cancer in US blacks and whites. Am J Epidemiol. 2002;155(11):1023–1032 [DOI] [PubMed] [Google Scholar]
- 17. Bosetti C, Talamini R, Montella M, et al. Retinol, carotenoids and the risk of prostate cancer: a case-control study from Italy. Int J Cancer. 2004;112(4):689–692 [DOI] [PubMed] [Google Scholar]
- 18. Key TJ, Appleby PN, Allen NE, et al. Plasma carotenoids, retinol, and tocopherols and the risk of prostate cancer in the European Prospective Investigation into Cancer and Nutrition study. Am J Clin Nutr. 2007;86(3):672–681 [DOI] [PubMed] [Google Scholar]
- 19. Pourmand G, Salem S, Mehrsai A, et al. The risk factors of prostate cancer: a multicentric case-control study in Iran. Asian Pac J Cancer Prev. 2007;8(3):422–428 [PubMed] [Google Scholar]
- 20. Kristal AR, Till C, Platz EA, et al. Serum lycopene concentration and prostate cancer risk: results from the Prostate Cancer Prevention Trial. Cancer Epidemiol Biomarkers Prev. 2011;20(4):638–646 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Schuurman AG, Goldbohm RA, Dorant E, et al. Vegetable and fruit consumption and prostate cancer risk: a cohort study in The Netherlands. Cancer Epidemiol Biomarkers Prev. 1998;7(8):673–680 [PubMed] [Google Scholar]
- 22. Peters U, Leitzmann MF, Chatterjee N, et al. Serum lycopene, other carotenoids, and prostate cancer risk: a nested case-control study in the prostate, lung, colorectal, and ovarian cancer screening trial. Cancer Epidemiol Biomarkers Prev. 2007;16(5):962–968 [DOI] [PubMed] [Google Scholar]
- 23. Beilby J, Ambrosini GL, Rossi E, et al. Serum levels of folate, lycopene, beta-carotene, retinol and vitamin E and prostate cancer risk. Eur J Clin Nutr. 2010;64(10):1235–1238 [DOI] [PubMed] [Google Scholar]
- 24. Huang HY, Alberg AJ, Norkus EP, et al. Prospective study of antioxidant micronutrients in the blood and the risk of developing prostate cancer. Am J Epidemiol. 2003;157(4):335–344 [DOI] [PubMed] [Google Scholar]
- 25. Jian L, Lee AH, Binns CW. Tea and lycopene protect against prostate cancer. Asia Pac J Clin Nutr. 2007;16(Suppl 1):453–457 [PubMed] [Google Scholar]
- 26. Etminan M, Takkouche B, Caamano-Isorna F. The role of tomato products and lycopene in the prevention of prostate cancer: a meta-analysis of observational studies. Cancer Epidemiol Biomarkers Prev. 2004;13(3):340–345 [PubMed] [Google Scholar]
- 27. Wu K, Erdman JW, Jr, Schwartz SJ, et al. Plasma and dietary carotenoids, and the risk of prostate cancer: a nested case-control study. Cancer Epidemiol Biomarkers Prev. 2004;13(2):260–269 [DOI] [PubMed] [Google Scholar]
- 28. Jian L, Du CJ, Lee AH, et al. Do dietary lycopene and other carotenoids protect against prostate cancer? Int J Cancer. 2005;113(6):1010–1014 [DOI] [PubMed] [Google Scholar]
- 29. Karppi J, Kurl S, Nurmi T, et al. Serum lycopene and the risk of cancer: the Kuopio Ischaemic Heart Disease Risk Factor (KIHD) study. Ann Epidemiol. 2009;19(7):512–518 [DOI] [PubMed] [Google Scholar]
- 30. Chug-Ahuja JK, Holden JM, Forman MR, et al. The development and application of a carotenoid database for fruits, vegetables, and selected multicomponent foods. J Am Diet Assoc. 1993;93(3):318–323 [DOI] [PubMed] [Google Scholar]
- 31. Mangels AR, Holden JM, Beecher GR, et al. Carotenoid content of fruits and vegetables: an evaluation of analytic data. J Am Diet Assoc. 1993;93(3):284–296 [DOI] [PubMed] [Google Scholar]
- 32. Tonucci LH, Holden JM, Beecher GR, Khachik F, Davis CS, Mulokozi G. Carotenoid content of thermally processed tomato-based food products. J Agric Food Chem. 1995;43(3):579–586 [Google Scholar]
- 33. Greene FL, American Joint Committee on Cancer, and American Cancer Society AJCC Cancer Staging Manual. 6th ed. New York: Springer-Verlag; 2002 [Google Scholar]
- 34. Mucci LA, Powolny A, Giovannucci E, et al. Prospective study of prostate tumor angiogenesis and cancer-specific mortality in the health professionals follow-up study. J Clin Oncol. 2009;27(33):5627–5633 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Pettersson A, Graff RE, Bauer SR, et al. The TMPRSS2:ERG rearrangement, ERG expression, and prostate cancer outcomes: a cohort study and meta-analysis. Cancer Epidemiol Biomarkers Prev. 2012;21(9):1497–1509 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Dhillon PK, Barry M, Stampfer MJ, et al. Aberrant cytoplasmic expression of p63 and prostate cancer mortality. Cancer Epidemiol Biomarkers Prev. 2009;18(2):595–600 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Zu K, Martin NE, Fiorentino M, et al. Protein expression of PTEN, insulin-like growth factor I receptor (IGF-IR), and lethal prostate cancer: a prospective study. Cancer Epidemiol Biomarkers Prev. 2013;22(11):1984–1993 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Willett W. Nutritional Epidemiology. 2nd ed. New York: Oxford University Press; 1998 [Google Scholar]
- 39. Kavanaugh CJ, Trumbo PR, Ellwood KC. The US Food and Drug Administration’s evidence-based review for qualified health claims: tomatoes, lycopene, and cancer. J Natl Cancer Inst. 2007;99(14):1074–1085 [DOI] [PubMed] [Google Scholar]
- 40. World Cancer Research Fund / American Institute for Cancer Research Food, Nutrition, Physical Activity, and the Prevention of Cancer: a Global Perspective. Washington DC: American Institute for Cancer Research; 2007 [Google Scholar]
- 41. Giovannucci E. A review of epidemiologic studies of tomatoes, lycopene, and prostate cancer. Exp Biol Med (Maywood). 2002;227(10):852–859 [DOI] [PubMed] [Google Scholar]
- 42. Unlu NZ, Bohn T, Clinton SK, et al. Carotenoid absorption from salad and salsa by humans is enhanced by the addition of avocado or avocado oil. J Nutr. 2005;135(3):431–436 [DOI] [PubMed] [Google Scholar]
- 43. Unlu NZ, Bohn T, Francis DM, et al. Lycopene from heat-induced cis-isomer-rich tomato sauce is more bioavailable than from all-trans-rich tomato sauce in human subjects. Br J Nutr. 2007;98(1):140–146 [DOI] [PubMed] [Google Scholar]
- 44. Giovannucci E. Commentary: Serum lycopene and prostate cancer progression: a re-consideration of findings from the prostate cancer prevention trial. Cancer Causes Control. 2011;22(7):1055–1059 [DOI] [PubMed] [Google Scholar]
- 45. Yang CM, Yen YT, Huang CS, et al. Growth inhibitory efficacy of lycopene and beta-carotene against androgen-independent prostate tumor cells xenografted in nude mice. Mol Nutr Food Res. 2011;55(4):606–612 [DOI] [PubMed] [Google Scholar]
- 46. Elgass S, Cooper A, Chopra M. Lycopene inhibits angiogenesis in human umbilical vein endothelial cells and rat aortic rings. Br J Nutr. 2012;108(3):431–439 [DOI] [PubMed] [Google Scholar]
- 47. Huang CS, Chuang CH, Lo TF, et al. Anti-angiogenic effects of lycopene through immunomodualtion of cytokine secretion in human peripheral blood mononuclear cells. J Nutr Biochem. 2012;24(2):428–434 [DOI] [PubMed] [Google Scholar]
- 48. Chen ML, Lin YH, Yang CM, et al. Lycopene inhibits angiogenesis both in vitro and in vivo by inhibiting MMP-2/uPA system through VEGFR2-mediated PI3K-Akt and ERK/p38 signaling pathways. Mol Nutr Food Res. 2012;56(6):889–899 [DOI] [PubMed] [Google Scholar]