Introduction

Micronutrients, antioxidants and vitamins have been shown to block initiation and suppress promotion and progression of cancer [1–3]. Studies have shown a reduced risk of prostate cancer with high intake and high serum levels of these micronutrients/antioxidants [3–9]. Intervention with tocopherol and selenium has been shown to reduce the incidence of prostate cancer in prospective randomised controlled trials [7, 9, 10]. Lycopene supplementation has been shown to reduce PSA and disease progression in localised prostate cancer patients when given preoperatively and to improve response to hormonal treatment [11–13]. Current prostate research suggests that the role of antioxidants and micronutrients may be greater in determining progression or stimulation of established tumours rather than in the development of new tumours [5, 11, 12, 14–17]

Prostate cancer is the commonest cancer in UK men, with over 27,000 cases diagnosed annually. There is extraordinarily wide inter-patient variation in the natural history of the disease, but accurate prediction of individual prostate cancer behaviour at the time of diagnosis is not currently possible. However, accurate prediction of individual prostate cancer behaviour at the time of diagnosis is not currently possible, and so immediate radical treatment for all cases is a standard approach. Curative intervention should ideally be targeted to the minority of men with significant cancers so that the remainder, with biologically irrelevant ‘disease’, are spared the risks of incontinence and impotence that are associated with treatment [18–20].

Active surveillance of early prostate cancer is an approach to managing localised prostate cancer in which the choice between curative treatment and observation is based on evidence of disease progression during a period of close monitoring [20–22]. The rate of rise of PSA shows significant inter-patient variation and is the major determinant of time to clinical progression [22–24]. Curative treatment is targeted to those men on active surveillance who have either a rapid PSA doubling time or adverse features on repeat biopsy [25, 26]. There is an unmet need for biomarkers of prostate cancer behaviour that could identify, at the time of diagnosis, those patients who could safely avoid radical treatment [20]. Active surveillance provides a uniquely valuable opportunity to study biomarkers of early prostate cancer behaviour [27]. If all prostate cancer patients receive immediate radical treatment, only 5–15% develop recurrence, which typically is detected years later; thousands of patients would need to be followed for many years for proper evaluation of candidate biomarkers. Furthermore, among the 85–95% that remain disease-free, it is impossible to distinguish the biologically irrelevant cancers from those that were significant but successfully treated. In contrast, outcome in terms of PSA kinetics and repeat biopsies is available for all men on active surveillance within a matter of months, so that candidate biomarkers can be evaluated rapidly, and in relatively few patients.

Biomarkers of prostate cancer behaviour could be important not just to avoid overtreatment of indolent disease but also to provide mechanistic insights into the extraordinary inter-patient variation seen in prostate cancer behaviour and thus inform the search for novel therapeutic targets. Serum micronutrient/antioxidant levels are attractive candidate markers of prostate cancer behaviour not just because of the evidence from epidemiological, in vitro and in vivo studies referred to above but also because if validated in patients by prospective randomised studies, they would provide an obvious opportunity for a low-toxicity intervention with the aim of preventing prostate cancer progression.

We have therefore studied the clinical outcomes in men with localised prostate cancer on active surveillance with respect to baseline serum levels of micronutrients/antioxidants.

Patients and methods

A prospective study of active surveillance of localised prostate cancer opened at the Royal Marsden Hospital in 2002. Eligibility criteria comprised histologically proven adenocarcinoma prostate, stage T1/2a N0/Nx with PSA levels <15 ng/ml, composite Gleason score ≤7 with primary grade ≤3 and percentage of positive biopsy cores ≤50% of total cores. All patients had monitoring of serum PSA every 3 months. PSA velocity was calculated using all the PSA values available from entry to the study until treatment or last follow-up, with a minimum of three consecutive PSA values a month apart. As part of the prospective protocol, all patients had repeat trans-rectal, ultrasound-guided, eight-core prostate biopsy at 18–24 months after entry to study. All pathology was verified centrally to avoid inter-observer variability. Radical treatment was recommended in the case of PSA progression defined as a PSA velocity >1 ng ml−1 year−1 or adverse histology on repeat biopsy defined as primary Gleason grade ≥4 or composite Gleason score more than 7% or >50% of the total cores positive for involvement. The study was approved by the Local Research Ethics Committee, and all patients gave written informed consent.

The active surveillance study protocol included banking of baseline serum samples for analysis of candidate prognostic biomarkers with respect to subsequent clinical outcomes. As part of an exploratory analysis, in a cohort of 104 consecutive patients, these baseline serum samples were analysed for levels of alpha-tocopherol, gamma-tocopherol, alpha-carotene and beta-carotene, lycopene, retinol and selenium using high-performance liquid chromatography/tandem mass spectrometry.

Statistical analysis

The main endpoint was time to disease progression, with disease progression defined as either histologic progression on repeat biopsy or radical treatment for PSA progression. Secondary endpoints were time to histologic disease progression and PSA velocity. Serum micronutrient/antioxidant levels were considered both as continuous variables and after dichotomising about the median. Univariate analysis of the relationship between time to progression or treatment and serum micronutrient/antioxidant levels was conducted by Cox regression method. The correlation of PSA velocity and the blood concentration of each micronutrient/antioxidant level was analysed using the Spearman rank correlation coefficient. Chi-squared test was utilised to analyse the difference between groups of patients who had adverse histology on repeat biopsy and the serum micronutrient/antioxidant levels dichotomised across the median, and a p value <0.05 was considered significant. Multivariate logistic regression with a backward stepwise selection process was utilised. Analysis was undertaken by SPSS 14 (SPSS Inc., IL).

Results

Blood samples from a cohort of 104 consecutive patients on the prospective active surveillance protocol were analysed. The patient characteristics are shown in Table 1. Median follow-up from the time of diagnosis was 2.5 years (95% CI 1–4 years). The median PSA velocity was 0.72 ng ml−1 year−1 (95% CI −1.47 to 3.33). Thirteen patients had treatment initiation on account of PSA progression. Twenty-five patients of the 91 patients who had a repeat biopsy had adverse histology on repeat biopsy. The median time to progression was 2.62 years (95% CI 2.2–3.03 years). None of the patients on active surveillance has developed distant metastasis or death from prostate cancer. The distribution of serum micronutrients/antioxidants levels in the study population are shown in Table 2. There was no correlation between baseline blood levels of alpha-tocopherol, gamma-tocopherol, alpha-carotene and beta-carotene, lycopene, retinol and selenium and clinical characteristics such as age, PSA, Gleason, PSA density or free/total PSA ratio (data not shown).

Table 1 Patient characteristics
Full size table
Table 2 Univariate analysis of baseline biochemical parameters to predict for time to progression or treatment by Cox regression and analysis of the correlation of blood micronutrient / antioxidant levels with PSA kinetics
Full size table

On univariate analysis, no significant association was seen between time to disease progression and serum levels of either alpha-tocopherol (p = 0.86), gamma-tocopherol (p = 0.84), alpha-carotenoid (p = 0.66), beta-carotene (p = 0.65), lycopene (p = 0.0.15), retinol (p = 0.76) or selenium (p = 0.76), as shown in Table 2. Multivariate regression analysis with a backward stepwise selection process did not produce any significant combination of variables to predict for disease progression. The analysis was repeated after accounting for confounding factors like age, initial PSA, etc. and also comparing the lowest and highest quartiles of individual serum values of the micronutrients and vitamins to predict for disease progression, which again did not yield any statistically significant results (data not shown).

No significant association was seen between adverse histology on repeat biopsy and serum levels of either alpha-tocopherol (p = 0.38), gamma-tocopherol (p = 0.98), alpha-carotenoid (p = 0.40), beta-carotene (p = 0.33), lycopene (p = 0.19), retinol (p = 0.55) or selenium (p = 0.70), as shown in Table 3. There was no significant correlation between PSA velocity and serum levels of either alpha-tocopherol (p = 0.22), gamma-tocopherol (p = 0.81), alpha-carotenoid (p = 0.99), beta-carotene (p = 0.97), lycopene (p = 0.97), retinol (p = 0.99) or selenium (p = 0.60), as shown in Table 2.

Table 3 Univariate analysis of dichotomised blood micronutrient and antioxidant levels to predict for adverse histology on repeat biopsy on repeat biopsy using the chi-squared test
Full size table

Discussion

From our prospective study of active surveillance, analysis of early clinical outcomes in a cohort of 104 consecutive patients with localised prostate cancer did not yield a significant correlation between the serum micronutrient/antioxidant levels and rate of disease progression. The hypothesis that blood levels of alpha-tocopherol, gamma-tocopherol, alpha-carotene and beta-carotene, lycopene, retinol and selenium are associated with the rate of disease progression in prostate cancer may require further investigation because if true, it could have important implications for the tertiary prevention of prostate cancer.

There is a considerable body of evidence from previous studies that prostate cancer patients have lower circulating concentrations of micronutrients and antioxidants such as lycopene, beta-carotene, alpha-tocopherol and retinol [6, 28]. Lycopene supplementation has lead to reduced PSA levels, has delayed or prevented progression of high-grade prostatic intraepithelial neoplasia to cancer and has improved the responses to hormonal treatment [13, 29, 30]. Kucuk et al. [11] have shown that lycopene supplementation before radical prostatectomy resulted in a reduction in PSA, reduction in tumour size, lesser involvement of surgical margins and extra-prostatic tissues with cancer and less diffuse high-grade prostatic intraepithelial neoplasia. Chang et al. [4] have suggested that higher circulating levels of alpha-carotene and trans-beta-carotene may contribute to lower prostate cancer risk. Weinstein et al. [9] reported from the results of the ATBC study that patients with a high-circulating alpha- and gamma-tocopherol levels had a lower risk of prostate cancer and that alpha-tocopherol supplementation reduced the incidence of prostate cancer by 32%. Prospective placebo-controlled clinical trial by Clark et al. in 974 patients have shown that supplementation with selenium resulted in a 63% reduction in the incidence of prostate cancer and a significant reduction in mortality from cancer [7, 10].

However, there is also evidence to the contrary like the SELECT study which showed no benefit from selenium and vitamin E in preventing prostate cancer in healthy men [17]. Similarly, the VITAL study did not show a significant correlation between the incidence of prostate cancer and the intake of vitamin E or selenium, though risk of clinically relevant advanced disease was reduced with greater long-term vitamin E supplementation [31]. In the large prospective Physician’s Health Study of Vitamin E and C supplementation, no appreciable reduction in prostate cancer or other malignancies were noted [32]. Plasma selenium concentration did not correlate with the risk of prostate cancer in a matched controlled European study [33].

Our study has several limitations. Though the blood levels of micronutrients/antioxidants may mirror intracellular levels of these chemicals, whether they are sufficient surrogate measurements to predict significant effect of biochemical parameters on disease progression is controversial. A snapshot of blood levels may also not be representative of or be predictive of the behaviour of prostate cancer considering the long natural history, with the wide fluctuations in the blood levels of these chemicals expected during this period. The lack of significant correlations between histologic progression and blood levels of micronutrients may also be due to tumour heterogeneity especially in biopsy specimens. The criteria for assessing progression in patients on active surveillance are not yet well defined, limitations of which include the use of PSA velocity as a ‘surrogate’ of long-term outcomes and the possibility of sampling error on repeat biopsy [25, 26]. Another drawback of our study is the short follow-up time compared to the long natural history of prostate cancer. It remains possible that blood levels of micronutrient and antioxidant levels in our cohort may be associated with important long-term outcomes such as time to metastasis, but to investigate this possibility will require many years of follow-up. Though our study may be underpowered due to low patient numbers, we could not detect any trend towards significance for any of the parameters analysed to predict for disease progression.

Supplementing early-stage prostate cancer patients with micronutrients/antioxidants, even in a study of short duration, has been shown to alter surrogate markers of proliferation such as serum PSA [13, 34]. Patients on active surveillance for localised prostate cancer may be ideal candidates for such supplementation, and it may be more effective in a population with low blood levels of these micronutrients/antioxidants. Prospective randomised controlled interventional studies are required to investigate and to assess the true preventive and therapeutic effects of these compounds and whether supplementation of these micronutrients/antioxidants, to produce adequate tissue concentration needed to induce suppression of tumour progression, would alter the biology of prostate cancer.