- Email: [email protected]
The epidemiology of prostate cancer
Urol Clin N Am 30 (2003) 209–217
The epidemiology of prostate cancer Peter Boyle, PhD, FRSEa,*, Gianluca Severi, PhDa, Graham G. Giles, PhDb a Division of Epidemiology and Biostatistics, European Institute of Oncology, Via Ripamonti 435, 20141 Milan, Italy Cancer Epidemiology Centre, The Cancer Council Victoria, 1 Rathdowne Street, Carlton South, Victoria 3053, Australia
b
Prostate cancer has become the most common cancer of males in several developed countries. With increasing age, most men will develop microscopic foci of prostate cancer whether they live in a population at a high or low risk for the invasive form of the disease [1] and, although a majority of men will develop microscopic disease, only a small percentage of these slow-growing tumors will develop into invasive prostate cancer and an even smaller proportion will cause premature death. The principal focus of epidemiologic investigations of prostate cancer, therefore, has to be the identification of factors— amenable to intervention—that cause the common microscopic form to progress to invasive disease. Prostate cancer is highest in Western populations, particularly among the black population of the United States. The disease is uncommon in many Asian and populations of developing countries [2]. Consideration of the public health importance of prostate cancer should be tempered
This study was conducted within the framework of support from the Associazione Italiana per la Ricerca sul Cancro (Italian Association for Research on Cancer) (P. Boyle, G. Severi) and the National Health and Medical Research Council (940394), and was further supported by funding from Tattersall’s and The Whitten Foundation as well as infrastructure provided by The Cancer Council of Victoria (G.G. Giles). Many of the ideas presented herein have been published previously and elsewhere by the authors. Thus, it is important to note that the authors do not claim this as an original work but merely a compilation of ideas brought together for review purposes. * Corresponding author. E-mail address: [email protected]
with the observation that in many countries the average age at death from prostate cancer is approaching 80 years (Fig. 1). Indeed, as data from England and Wales demonstrate, the average age at diagnosis of prostate cancer is greater by a considerable margin than is the average age at diagnosis for other common cancers such as breast and colorectal. The epidemiology of prostate cancer has been notoriously difficult to study and the disease continues to present formidable challenges to epidemiologists. Much of the difficulty is linked with the lack of knowledge regarding disease specificity. Both the phenotypes and genotypes of prostate cancers are heterogeneous and studies that combine all forms of prostate cancer together are, therefore, likely to attenuate any associations that might only arise with particular subtypes. This problem has gained more widespread recognition in recent years, and epidemiologists have attempted to increase disease specificity in their studies, largely by stratifying on severity—for example, histologic grade, Gleason scores, stage of disease, progression, and death. Although this approach has occasionally produced strengthened associations, it has not greatly advanced etiologic understanding. The causes of prostate cancer have been investigated in numerous case-control studies and a few prospective cohort studies [3–5]. In addition to disease specificity, there have been other problems with epidemiologic studies of prostate cancer, particularly small sample sizes and, therefore, poor statistical power, poor exposure measurement, and inappropriate study designs. The best available epidemiologic evidence on prostate cancer can be obtained from only a handful of large well-controlled case-control
0094-0143/03/$ - see front matter Ó 2003, Elsevier Science (USA). All rights reserved. doi:10.1016/S0094-0143(02)00181-7
210
P. Boyle et al / Urol Clin N Am 30 (2003) 209–217
Fig. 1. Cumulative distribution of numbers of deaths from cancer in the United Kingdom, 1990.
studies and cohort studies. Although historically, case-control studies have identified numerous putative risk factors, only age and a family history of prostate cancer are well established. During the 1990s, prospective studies suggested that specific fatty acids, antioxidant vitamins, and carotenoids might alter prostate cancer risk. There also were reports that changes in plasma levels of key hormones and associated molecules, and naturally occurring variants in genes (polymorphisms) of the androgen, vitamin D, and insulin-like growth factor 1 (IGF-1) prostate cell growth regulatory pathways also might alter risk, and conjectures that dietary factors might modulate risk by interacting with these pathways. Population trends Prostate cancer is extremely age dependent, is uncommon before the age of 50, and increases rapidly thereafter. It demonstrates a large variation in incidence between countries. For example, the age-standardized incidence rates per 100,000 from the seventh edition of Cancer Incidence in Five Continents [2], which covers the period 1988–1992, ranged from very high in the United States (data derived from the Surveillance, Epidemiology and End Results (SEER) database of the National Cancer Institute: SEER blacks, 137; SEER whites, 101), to much lower in Europe (England and Wales, 28), to low in Asia (Japan, Miyagi Registry, 9). There are also substantial differences between ethnic groups living in the same country; for example, Japanese Americans have rates that are intermediate to those of the SEER whites and native Japanese stated above (Hawaii, 64; Los Angeles, 47) [2]. The changes in incidence observed in migrant populations are of interest because they point to factors that may increase prostate cancer risk and also to those that may reduce it.
Prostate cancer has increased in incidence in many countries around the world. In aging populations, the burden of prostate cancer can be expected to increase over time, even if the agespecific rates remain constant; however, between the mid 1980s and the early 1990s incidence rose dramatically, especially in parts of the United States [6]. The substantial changes in prostate cancer incidence and mortality that have occurred in the United States since the introduction of the prostate-specific antigen (PSA) blood test are of enormous interest and importance. Trends can be examined from 1975 [6]. Between 1986 and 1992, the overall age-adjusted incidence rate more than doubled from 86 to 179 per 100,000 per annum. Following this rise, incidence declined in all categories of age considered (under 65 years, 65– 74 years, and 75 years and older). The increase was most apparent in younger men (up to age 65), whereas the recent decline was most notable in men aged 75 and older. The increase in incidence between 1986 and 1992 occurred for both localized and regional stages of prostate cancer and was greatest in moderately differentiated tumors. The incidence of distant stage prostate cancer peaked in 1985 and has since declined [7]. From the pre-PSA era (1980–1985) until the PSA-era (1990–1995), the median age at diagnosis in the United States decreased by 1 year for whites and blacks. During the same period, the median age at death increased by 1 year in each racial group [7]. The median age at diagnosis in the United States (71 years), however, was considerably less than that in England and Wales (77 years) for the period 1988–1990 [8]. Undoubtedly, the trends of cancer incidence in the United States have been influenced by the widespread use of PSA testing. It may be premature, however, to attribute the recent reduction in prostate cancer mortality to PSA testing [9,10]. First, the reduction in mortality appeared to occur quickly to be confidently ascribed to screening and, second, the introduction of PSA was contemporaneous with that of widespread androgen deprivation therapy, a treatment that has had a significant impact on mortality. Family history and genetics Prostate cancer demonstrates a pattern of familial aggregation similar to that shown by breast and colorectal cancers, with an increased risk to a first-degree relative of a man with prostate cancer of about 2 to 3 on average [11–13],
P. Boyle et al / Urol Clin N Am 30 (2003) 209–217
the risk increasing the younger the age at diagnosis of the affected member. Even if a risk of 1.5 associated with having an affected first-degree relative was due to genetic factors, the combined effects of those genetic factors would have a large effect on disease risk, equivalent to an interquartile risk ratio an order of magnitude greater [14]. Thus, a similar degree of familial aggregation can be a consequence of either a rare high-risk genetic mutation or a common low-risk genetic polymorphism. Interest in the role of genetic factors in prostate cancer susceptibility has escalated due to a number of linkage analyses based on families that contain several men with prostate cancer, most of whom have early-onset disease. As discussed in a recent review [13], convincing replications have been rare, and heterogeneity analyses suggest that if any one of these regions contains a major prostate cancer gene, mutations in that gene would explain only a small proportion of multiple-case prostate cancer families. On the other hand, there have been reports of more modest risks of prostate cancer associated with common variants (polymorphisms) in candidate genes, such as those that encode the androgen receptor (AR), PSA, 5a-reductase type 2 (SRD5A2), cytochrome P450, vitamin D receptor (VDR), glutathione S-transferase, and HPC2/ELAC2 [15–21]. Thus far, early reports of associations have not been replicated in later studies—a pattern consistent with ‘‘data torture’’ of small clinical series. If true, however, the modest risks associated with common polymorphisms might explain a far greater proportion of disease than do the high risks associated with rare mutations. Hormones and other growth factors Growth and maintenance of normal prostate epithelium is regulated by the androgen and vitamin D pathways, with androgens stimulating and vitamin D metabolites inhibiting cell proliferation [22]. The androgen and vitamin D pathways interact at various levels, with one endpoint of both being the IGF-1 axis [22]. Perturbations of all three pathways have been associated with prostate cancer [22]. Cell division in the prostate is controlled by testosterone (T) that is reduced in the prostate to its active form, 5a-dihydrotestosterone (DHT), by 5-alpha reductase (SRD5A2) [23]. DHT binds to the AR to induce dimerization and binding to androgen response elements of
211
target genes to regulate their transcription [24]. Allelic variants of SRD5A2 (49T, 89V), which are thought to increase SRD5A2 activity have been associated with an increased risk of prostate cancer [25]. Short alleles of the AR CAG microsatellite also have been associated with increased risk and with cancers of aggressive phenotype [25]. AR genes with short CAG regions are more highly expressed compared with AR genes with longer CAG regions [26,27]. Other polymorphisms have been identified in genes encoding androgen biosynthetic and catabolic enzymes (such as CYP17, HSD3B2, and HSD17B3), but their association with prostate cancer has not been determined [26,28]. Consistent with the hypothesis that increased AR activity increases the risk of prostate cancer, prospective risk studies of androgen plasma/serum measurements suggest that a high plasma T to DHT ratio, high circulating levels of T, low levels of dehydroepiandrosterone or low levels of sex hormone binding globulin (SHBG)—which binds to T, decreasing its bioavailability—may elevate risk. Whereas androgens stimulate prostate cell proliferation, vitamin D inhibits it [29]. Two small nested case-control studies [30,31] have reported that high levels of the vitamin D metabolite 1,25D in prediagnostic sera are associated with lower risk of prostate cancer. Thus, alterations in the VDR gene that affect receptor activity might affect prostate cancer susceptibility. Three restriction fragment length polymorphisms as well as a polymorphism in the translation initiation site of the VDR gene, and a poly A length polymorphism have been identified [32,33]. It is not clear whether any of these affect VDR function, and although preliminary studies [33,34] implicate the poly A and the Taq I polymorphisms in prostate cancer risk, they need to be replicated. Both the AR and the VDR are thought to produce some of their growth effects via IGF-1. Binding proteins regulate IGF-1 availability in the prostate [22]. Expression of the major binding protein (IGF-BP3) is regulated by androgens and vitamin D; androgens inhibit expression of IGFBP3 [24], whereas an analogue of vitamin D has been shown to upregulate expression of IGF-BP3 [33]. The association between circulating IGF-1 levels and prostate cancer risk [3,4] was confirmed in a prospective study [4] that showed an approximate doubling of risk per 100-ng/mL increase in IGF-1, in serum samples collected before subsequent development of prostate cancer. This association strengthened when controlling for
212
P. Boyle et al / Urol Clin N Am 30 (2003) 209–217
IGF-BP3 [4]. Although a number of polymorphisms upstream of the IGF-1 transcriptional start site have been identified [34,35], direct association of these polymorphisms with prostate cancer risk has yet to be evaluated. Whites who are homozygous for the (CA) 19 alleles have significantly lower IGF-1 serum levels than do other genotypes [36]. Environmental factors Meat and fat The evidence of an association between meat and prostate cancer is more consistent than that for fat [37], particularly for advanced disease. The association with meat intake is difficult to separate from that of fat, however, because meat is a major vehicle for fat in Western diets. There is certainly insufficient evidence on which to base dietary advice. One possible explanation for the limited evidence is that meat produces carcinogenic heterocyclic amines when cooked at high temperature, and also produces polycyclic aromatic hydrocarbons when grilled over flames. The only study on this topic [38] has produced equivocal findings. Cohort studies [3,4] have reported positive associations between prostate cancer and red meat consumption, total animal fat consumption, and intake of fatty animal foods. It is thought that dietary fatty acids may affect risk in a variety of ways, including altering serum levels of sex hormones or modulating eicosanoid synthesis and affecting tumor cell proliferation, immune response, invasion, and metastasis; altering the composition of cell membrane phospholipids; affecting 5a-reductase type I activity [39]; forming free radicals from peroxidation; decreasing 1,25 D levels; and increasing IGF-1 levels. Evidence suggests that increased biosynthesis by prostate cancer cells of arachidonic acid-derived prostaglandins and hydroxyacid eicosanoids via the cycloxegenase type 2 (COX-2) enzyme pathway results in enhanced cancer cell proliferation and invasive and metastatic behavior. The COX enzyme may be inhibited directly by nonsteroidal anti-inflammatory drugs [40–42]. Plant foods, antioxidant vitamins, carotenoids, and minerals In general, vegetable consumption has not been strongly associated with prostate cancer risk [43], with the exception of a protective affect of tomatoes and foods that contain concentrated
tomato products. Lycopene, a fat-soluble carotenoid principally found in tomatoes, is an efficient singlet oxygen quencher, and has been shown to be unusually concentrated in the prostate gland [44]. In one cohort study [3,4], high lycopene intake was related to a 21% lower risk of prostate cancer (P<0.05), especially for advanced disease. In another study, the risk in men with higher plasma lycopene levels was reduced by 25%, and, for aggressive cancer, was reduced by 44%. Some studies support a possible protective effect of legumes, carrots, and green leafy and cruciferous vegetables, but this data is not compelling, nor is the evidence for other carotenoids. Vitamin A has been reported to both increase and reduce the risk of prostate cancer. Similarly, there is mixed evidence with regard to the effects of dietary beta-carotene: although some casecontrol studies have suggested a protective effect, no benefit was seen in large prospective studies. The evidence of a protective effect of vitamin E (a-tocopherol) supplements is more convincing. Male smokers randomized to take 50-mg a-tocopherol supplement in the Alpha-Tocopherol BetaCarotene (ATBC) Trial [45], had a statistically significant 32% decrease in clinical prostate cancer incidence, and a significant 41% reduction in prostate cancer mortality, compared with the placebo group. This should be interpreted with particular caution as this was not among the primary hypotheses for the study. Evidence of an effect associated with the amounts of a-tocopherol that can be obtained from dietary sources is weak. Another study [46] suggested that a-tocopherol is more potent and interacts with c-tocopherol. The consistent positive association between prostate cancer and dairy products that are rich in vitamin D may be explained by the high calcium content in dairy foods, which suppresses the formation and circulation of 1,25D levels. Two studies [4] have found strong positive associations between calcium intake and prostate cancer. Therefore, associations between fructose intake (negative) and meat intake (positive) and prostate cancer risk could be explained, in part, by effects on 1,25D levels. Phytoestrogens include two groups of hormonelike diphenolic compounds—isoflavonoids and lignans. Their consumption has been proposed as a contributing factor to the low levels of prostate and breast cancers seen in societies that consume high levels of soy products and other legumes. Phytoestrogens have been shown to act as antioestrogens, weak estrogens, and antioxi-
P. Boyle et al / Urol Clin N Am 30 (2003) 209–217
dants, and have been shown to inhibit prostate tumor model systems in vivo and in vitro [47]. Although their biologic function is not understood fully, they have been reported to inhibit 5a-reductase types 1 and 2 and 17b-hydroxysteroid dehydrogenase and aromatase, and to stimulate the synthesis of SHBG and uridine diphosphate glucuronyltransferase, suggesting that they may act, in part, by decreasing the biologically available fraction of androgens [47]. Very little analytical epidemiologic research is available [48]. Similarly, polyphenolic compounds in green tea, called catechins, are believed to have risk-reducing properties, but the epidemiologic evidence is scarce and inconsistent [49]. Selenium and zinc both have received attention as possible chemopreventive agents for prostate cancer, so much so that selenium (and vitamin E) supplementation is being tested in a randomized trial. The prostate has a unique relationship with zinc metabolism that is related to the zinc content of seminal fluid. A case-control study [50] has reported that zinc supplement use is associated in a significant dose–response fashion with a reduced risk of prostate cancer. Energy intake, energy balance, and body composition Energy intake may be associated with prostate cancer risk via an effect on IGF-1 levels [51]. Energy imbalance leads to adiposity. Two thirds of plasma E in men is converted from dehydroepiandrosterone and androstenedione by aromatase in adipose tissue and muscle. Thus, body composition affects the ratio of circulating E and T. Prostate cells are sensitive to T and E because they express both the a and b forms of the E receptor [52]. Continuous long-term exposure to a low level of E in combination with androgen during adulthood induces prostate enlargement in mice, rats, and dogs [53]. These findings suggest that an increased exposure to E at any time in life can affect prostate cell growth and prostate growth regulatory dysfunction. Associations between body mass index (BMI) and prostate cancer have been weak and inconsistent, perhaps reflecting the inadequacy of BMI as a measure of body composition. Physical activity theoretically could reduce prostate cancer risk by reducing plasma T levels. The evidence from epidemiologic studies is inconsistent, but suggestive of a protective effect, possibly restricted to high physical activity levels. For example, two recent prospective cohort
213
studies in the United States [51,52] showed no evidence of an association with physical activity, whereas another [54] showed that physically inactive men were at increased risk compared with very active men. This latter association, however, was limited to black Americans and was not statistically significant in whites. A cohort study of 22,895 Norwegian men gave a relative risk (RR) of 0.8 (confidence interval [CI]=0.62–1.03) for high versus low activity [55]. Physical activity is negatively correlated with obesity, and yet both are negatively correlated with T levels. There is only inconsistent evidence that obesity (as measured by BMI) is associated with prostate cancer. BMI is a problematic measure in this regard because it combines both adiposity and lean body mass—the two components having different hormonal associations, with the latter being under the influence of androgens and IGF-1. This issue will be resolved only by cohort studies that include separate estimates of lean and fat body mass [56]. Sexual factors Given that prostate growth and development is androgen dependent, several studies [5,57–60] have pursued the idea that prostate cancer risk might be related to variation in androgen milieu manifested by differences in sexual activity; however, these studies have reported only weak and inconsistent findings. One focus has been sexually transmitted diseases (STDs) and behaviors that are associated with increased risk of infection, such as multiple sex partners and sex with prostitutes. Apart from a history of STDs, there is limited evidence that papillomavirus infection increases risk [61]. The literature also is inconsistent with regard to sexual activity. Rotkin’s [62] proposal that reduced ejaculatory frequency in normal men increases the risk of prostate cancer is supported by a casecontrol study [63], which reported an odds ratio of 4.05 (CI=2.99–5.48) for men who had experienced an (undefined) period of interrupted sexual activity. Further support for Rotkin’s theory comes from a study that showed that presumably celibate Roman Catholic priests had an above-average risk of dying from prostate cancer [64]. Tobacco and alcohol Smoking is most likely not associated with prostate cancer incidence; however, there is some evidence that smoking may be positively associated with mortality from this cancer [65,66].
214
P. Boyle et al / Urol Clin N Am 30 (2003) 209–217
Recent cohort studies [67,68] gave estimates that were close to unity for associations between the incidence of all prostate cancer and either current or past smoking and modest positive associations cancer (RRs between 1.3 and 2) with respect to fatal prostate. A competent review [69] concluded that there was no association between low to moderate alcohol consumption and prostate cancer, but could not exclude the possibility of an association with heavy drinking or that the effects of alcohol might be observed in subgroups defined by genetic markers and family history. More recent findings [70–72] did not modify this conclusion. Occupation The accumulated studies of occupation have failed to provide compelling evidence of associations with the risk of prostate cancer. The weak associations in the literature are probably the result of poorly measured exposures and uncontrolled confounding with social class. Cadmium exposure once was highly suspected, but further follow-up of industrial cohorts failed to confirm evidence of an association. A number of studies have reported associations with farming—with particular reference to pesticides, but most have not measured exposures at the level of the individual. Summary The etiology of prostate cancer remains virtually unknown. Although there are a number of new leads with regard to risk factors for prostate cancer, more research is required to confirm them. There is little purpose in conducting further casecontrol studies of prostate cancer—particularly since the use of PSA testing has become widespread. Instead, future epidemiologic studies should focus on prostate tumor subclassification, in terms of method of detection, markers of biological ‘‘aggressiveness,’’ and genetic changes. Many of these new leads involve the possible influence of polymorphisms in key genes involved in important physiologic processes in the prostate. To fully explore the complexity of interrelationships between the several elements in these pathways will require large cohort studies in which blood is sampled prior to diagnosis. Such studies will be important for identifying which modifiable aspects of lifestyle (such as diet, alcohol, tobacco, physical activity) might be targeted for intervention, to reduce risk.
The detection of early prostate cancers by PSA testing relatives of men with prostate cancer has affected the prevalence of phenocopies and, hence, the meaningfulness of risk estimation in prostate cancer families. Because multiple-case families form the substrate for gene hunting via linkage analysis, this phenocopy phenomenon is going to cause considerable confusion and wasted effort. Presently, men with a family history of prostate cancer can be provided with little advice in terms of preventive action. It is likely that one or more genetic mutations associated with a high risk for prostate cancer will be identified in the near future. Even so, the risks probably will be similar to those for mutations in the first two breast cancer genes—informative for very few families. It is difficult to foresee, as and when high-risk mutation carriers are identified, what advice should be offered to them: prophylactic prostatectomies seem to have less attraction than do prophylactic mastectomies for women at high risk of breast cancer. This issue becomes more complex when considering counseling on the basis of a genetic profile involving many low-risk polymorphisms. Hopefully, such genetic screening should occur only after its efficacy has been established; when there is a better understanding of prostate biology, tumor heterogeneity, and prognosis; and when there are proven treatment or prevention options available. Prevention is held to be better than cure, and there are some potentially interesting chemopreventive agents that require careful trial before public health initiatives can be promoted. Potential candidates include vitamin E, selenium, zinc, and lycopene as dietary supplements. There are other agents that may be appropriate for pharmaceutical development, including inhibitors of COX-2 and IGF-1 activity. It is important that chemoprevention trials are followed-up for a sufficient period of time and that other endpoints also are captured, because the supplementation of diets with superphysiologic doses of individual micronutrients sometimes has caused unexpected and unwanted results—for example, an 18% increase in lung cancers observed in the beta-carotene arm of the ATBC trial [73].
References [1] Yatani R, Chigusa I, Akazaki K, et al. Geographic pathology of latent prostatic carcinoma. Int J Cancer 1982;29:611–6.
P. Boyle et al / Urol Clin N Am 30 (2003) 209–217 [2] Parkin DM, Whelan SL, Ferlay J, Raymond L, Young J, editors. Cancer incidence in five continents, Vol 7, Lyon, France; International Agency for Research on Cancer; 1997. [3] Clinton SK, Giovannucci E. Diet, nutrition, and prostate cancer. Annu Rev Nutr 1998;18:413–40. [4] Chan JM, Stampfer MJ, Giovannucci EL. What causes prostate cancer? A brief summary of the epidemiology. Semin Cancer Biol 1998;8:263–73. [5] Ross RK, Schottenfeld D. Prostate cancer. In: Schottenfeld D, Fraumeni JF Jr, editors. Cancer epidemiology and prevention. 2nd edition. New York: Oxford University Press; 1996. p. 1180–206. [6] Hsing AW, Devesa SS. Trends and patterns of prostate cancer: what do they suggest? Epidemiol Rev 2001;23(1):3–13. [7] Newcomer LM, Stanford JL, Blumenstein BA, et al. Temporal trends in rates of prostate cancer: declining incidence of advanced stage disease, 1974 to 1994. J Urol 1997;158:1127–30. [8] Boyle P, Severi G. Epidemiology of prostate cancer chemoprevention. Eur Urol 1999;35(5–6):370–6. [9] Tarone RE, Chu KC, Brawley O. Implications of stage-specific survival rates in assessing recent declines in prostate cancer mortality rates. Epidemiology 2000;11(2):167–70. [10] Oliver SE, Gunnell D, Donovan JL. Comparison of trends in prostate-cancer mortality in England and Wales and the USA. Lancet 2000;355(9217):1788–9. [11] Lesko SM, Rosenberg L, Shapiro S. Family history and prostate cancer risk. Am J Epidemiol 1996; 144:1041–7. [12] Hopper JL, Carlin JB. Familial aggregation of a disease consequent upon correlation between relatives in a risk factor measured on a continuous scale. Am J Epidemiol 1992;136:1138–47. [13] Ostrander EA, Stanford JL. Genetics of prostate cancer: too many loci, too few genes. Am J Hum Genet 2000;67:1367–75. [14] Peto J. Genetic predisposition to cancer. In: Cairns J, Lyon JL, Skolnick M, editors. Cancer incidence in defined populations (Banbury Report no. 4). Cold Spring Harbor (NY): Cold Spring Harbor Laboratory Press; 1980. p. 203–13. [15] Ingles SA, Ross RK, Yu MC, et al. Association of prostate cancer risk with genetic polymorphisms in vitamin D receptor and androgen receptor. J Natl Cancer Inst 1997;89:166–70. [16] Xue W, Irvine RA, Yu MC, et al. Susceptibility to prostate cancer: interaction between genotypes at the androgen receptor and prostate specific antigen loci. Cancer Res 2000;60:839–41. [17] Jaffe JM, Malkowicz SB, Walker AH, et al. Association of SRD5A2 genotype and pathological characteristics of prostate tumors. Cancer Res 2000; 60:1626–30. [18] Walker AH, Jaffe JM, Gunasegaram S, et al. Characterization of an allelic variant in the nifedipine-specific element of CYP3A4: ethnic distribu-
[19]
[20]
[21]
[22]
[23] [24] [25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
215
tion and implications for prostate cancer risk. Hum Mutat 1998;12:289. Ma J, Stampfer MJ, Gann PH, et al. Vitamin D receptor polymorphisms, circulating vitamin D metabolites, and risk of prostate cancer in US physicians. Cancer Epidemiol Biomarkers Prev 1998;242:467–73. Kelada SN, Kardia SL, Walker AH, et al. The glutathione S-transferase-mu and -theta genotypes in the etiology of prostate cancer: genotype– environment interactions with smoking. Cancer Epidemiol Biomarkers Prev 2000;9:1329–34. Rebbeck TR, Walker AH, Zeigler-Johnson C, et al. Association of HPC2/ELAC2 genotypes and prostate cancer. Am J Hum Genet 2000;67:1014–9. Russell PJ, Bennett S, Stricker P. Growth factor involvement in progression of prostate cancer. Clin Chem 1998;44:705–23. Santen RJ. Endocrine treatment of prostate cancer. J Clin Endocrinol Metab 1992;75:685–9. Bentel JM, Tilley WD. Androgen receptors in prostate cancer. J Endocrinol 1996;151:1–11. Ross RK, Pike MC, Coetzee GA, et al. Androgen metabolism and prostate cancer: establishing a model of genetic susceptibility. Cancer Res 1998; 58:4497–504. Choong CS, Kemppainen JA, Zhou Z-X, et al. Reduced androgen receptor gene expression with first exon CAG repeat expansion. Mol Endocrinol 1996;10:1527–35. Chamberlain NL, Driver ED, Miesfeld RL. The length and location of CAG trinucleotide repeats in the androgen receptor N-terminal domain affect transactivation function. Nucleic Acids Res 1994;22:3181–6. Moghrabi N, Hughes E, Dunaif A, et al. Deleterious missense mutations and silent polymorphism in the human 17b-hydroxysteroid dehydrogenase 3 gene (HSD17B3). J Clin Endocrinol Metab 1998; 83:2855–60. Issa LL, Leong GM, Eisman JA. Molecular mechanism of vitamin D receptor action. Inflamm Res 1998;47:451–75. Ingles SA, Ross RK, Yu MC, et al. Association of prostate cancer risk with genetic polymorphisms in vitamin D receptor and androgen receptor. J Natl Cancer Inst 1997;89:166–71. Taylor JA, Hirvonen A, Watson M, et al. Association of prostate cancer with vitamin D receptorgene polymorphism. Cancer Res 1996;56: 4108–4110. Corder EH, Friedman GD, Vogelman JH, et al. Seasonal variation in vitamin D, vitamin D-binding protein, and dehydroepiandrosterone: risk of prostate cancer in black and white men. Cancer Epidemiol Biol Prev 1995;4:655–9. Nickerson T, Huynh H. Vitamin D analogue EB1089-induced prostate regression is associated with increased gene expression of insulin-like
216
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42] [43]
[44]
[45]
[46]
[47]
[48]
[49] [50]
P. Boyle et al / Urol Clin N Am 30 (2003) 209–217 growth factor binding proteins. J Endocrinol 1999; 160:223–9. Miyao M, Hosoi T, Inoue S, et al. Polymorphism of insulin-like growth factor I gene and bone mineral density. Calcif Tissue Int 1998;63:306–11. Mullis PE, Patel MS, Brickell PM, et al. Constitutionally short stature: analysis of the insulin-like growth factor-I gene and the human growth hormone gene cluster. Pediatr Res 1991;29:412–5. Rosen CJ, Kurland ES, Vereault D, et al. Association between IGF I and a simple sequence repeat in the IGF I gene: implications for genetic studies of bone mineral density. J Clin Endocrinol Metab 1998;83:2286–90. Kolonel LN, Nomura AM, Cooney RV. Dietary fat and prostate cancer: current status. J Natl Cancer Inst 1999;91(5):414–28. Norrish AE, Ferguson LR, Knize MG, et al. Heterocyclic amine content of cooked meat and risk of prostate cancer. J Natl Cancer Inst 1999; 91:2038–44. Liang T, Liao S. Inhibition of steroid 5a-reductase by specific aliphatic unsaturated fatty acids. Biochem J 1992;285:557–62. Zhou JR, Blackburn GL. Bridging animal and human studies: what are the missing segments in dietary fat and prostate cancer? Am J Clin Nutr 1997;66:1572S–80S. Norrish AE, Jackson RT, McRae CU. Nonsteroidal anti-inflammatory drugs and prostate cancer progression. Int J Cancer 1998;77:511–5. Ghosii J, Myers C. Arachidonic acid metabolism and cancer of the prostate. Nutrition 1998;14:48–9. Chan JM, Giovannucci EL. Vegetables, fruits, associated micronutrients, and risk of prostate cancer. Epidemiol Rev 2001;23:82–6. Clinton SK, Emenhiser C, Schwartz SJ, et al. Cistrans lycopene isomers, carotenoids, and retinol in the human prostate cancer. Cancer Epidemiol Biomarkers Prev 1996;5:823–33. Rautalahti M, Albanes D, Virtamo J, et al. Betacarotene did not work: aftermath of the ATBC study. Cancer Lett 1997;114:235–6. Helzlsouer KJ, Huang HY, Alberg AJ, Hoffman S, Burke A, Norkus EP, et al. Association between alpha-tocopherol, gamma-tocopherol, selenium, and subsequent prostate cancer. J Natl Cancer Inst 2000;92(24):2018–23. Griffiths K, Denis L, Turkes A, Morton MS. Possible relationship between dietary factors and pathogenesis of prostate cancer. Int J Urol 1998; 5:195–213. Barnes S. Role of phytochemicals in prevention and treatment of prostate cancer. Epidemiol Rev 2001; 23(1):102–5. Chhabra SK, Yang CS. Tea and prostate cancer. Epidemiol Rev 2001;23:106–9. Kristal AR, Stanford JL, Cohen JH, et al. Vitamin and mineral supplement use is associated with
[51]
[52]
[53]
[54]
[55]
[56]
[57]
[58]
[59]
[60]
[61]
[62]
[63]
[64]
[65]
[66]
reduced risk of prostate cancer. Cancer Epidemiol Biomarkers Prev 1999;8:887–92. Putnam SD, Cerhan JR, Parker AS, et al. Lifestyle and anthropometric risk factors for prostate cancer in a cohort of Iowa men. Ann Epidemiol 2000; 10:361–9. Liu SE, Lee IM, Linson P, et al. A prospective study of physical activity and risk of prostate cancer in US physicians. J Epidemiol 2000;29:29–35. vom Saal FS, Timms BG, Montano MM, et al. Prostate enlargement in mice due to fetal exposure to low doses of estradiol or diethylstilbestrol and opposite effects at high doses. Proc Nat Acad Sci 1997;94:2056–61. Clarke G, Whittemore AS. Prostate cancer risk in relation to anthropometry and physical activity: the National Health and Nutrition Examination Survey Epidemiological Follow-Up Study. Cancer Epidemiol Biomarkers Prev 2000;9:875–81. Lund Nilsen TI, Johnsen R, Vatten LJ. Socioeconomic and lifestyle factors associated with the risk of prostate cancer. Br J Cancer 2000;82: 1358–1363. Severson RK, Grove JS, Nomura AMY, Stemmerman GN. Body mass and prostate cancer: a prospective study. BMJ 1988;297:713–5. Ewings P, Bowie C. A case-control study of cancer of the prostate in Somerset and east Devon. Br J Cancer 1996;74(4):661–6. Hayes RB, Pottern LM, Strickler H, et al. Sexual behaviour, STDs and risks for prostate cancer. Br J Cancer 2000;82(3):718–25. Andersson SO, Baron J, Bergstrom R, Lindgren C, Wolk A, Adami HO. Lifestyle factors and prostate cancer risk: a case-control study in Sweden. Cancer Epidemiol Biomarkers Prev 1996;5(7):509–13. Hsieh CC, Thanos A, Mitropoulos D, Deliveliotis C, Mantzoros CS, Trichopoulos D. Risk factors for prostate cancer: a case-control study in Greece. Int J Cancer 1999;80(5):699–703. Strickler HD, Goedert JJ. Sexual behavior and evidence for an infectious cause of prostate cancer. Epidemiol Rev 2001;23(1):144–51. Rotkin ID. Studies in the epidemiology of prostatic cancer: expanded sampling. Cancer Treat Rep 1977;61(2):173–80. Fincham SM, Hill GB, Hanson J, Wijayasinghe C. Epidemiology of prostatic cancer: a case-control study. Prostate 1990;17(3):189–206. Ross RK, Deapen DM, Casagrande JT, et al. A cohort study of mortality from cancer of the prostate in Catholic priests. Br J Cancer 1981; 43:233–5. Colditz G. Consensus conference: smoking and prostate cancer. Cancer Causes Control 1996;7: 560–2. Lumey LH. Prostate cancer and smoking: a review of case-control and cohort studies. Prostate 1996; 29:249–60.
P. Boyle et al / Urol Clin N Am 30 (2003) 209–217 [67] Giovannucci E, Rimm EB, Ascherio A, et al. Smoking and risk of total and fatal prostate cancer in United States Health professionals. Cancer Epidemiol Biomarkers Prev 1999;8:277–82. [68] Lotufo PA, Lee IM, Ajani UA, et al. Cigarette smoking and risk of prostate cancer in the Physicians’ Health Study. Int J Cancer 2000;87: 141–4. [69] Breslow RA, Weed DL. Review of epidemiologic studies of alcohol and prostate cancer: 1971–1996. Nutr Cancer 1998;30:1–13. [70] Lumey LH, Pittman B, Wynder EL. Alcohol use and prostate cancer in U.S. whites: no association in a confirmatory study. Prostate 1998;36:250–55.
217
[71] Villeneuve PJ, Johnson KC, Krieger N, Mao Y. Risk factors for prostate cancer: results from the Canadian National Enhanced Cancer Surveillance System. The Canadian Cancer Registries Epidemiology Research Group. Cancer Causes Control 1999;10:355–67. [72] Schuurman AG, Goldbohm RA, van den Brandt PA. A prospective study on consumption of alcoholic beverages in relation to prostate cancer incidence (The Netherlands). Cancer Causes Control 1999;10:597–605. [73] Rautalahti M, Albanes D, Virtamo J, et al. Betacarotene did not work: aftermath of the ATBC study. Cancer Lett 1997;114:235–6.
The epidemiology of prostate cancer Peter Boyle, PhD, FRSEa,*, Gianluca Severi, PhDa, Graham G. Giles, PhDb a Division of Epidemiology and Biostatistics, European Institute of Oncology, Via Ripamonti 435, 20141 Milan, Italy Cancer Epidemiology Centre, The Cancer Council Victoria, 1 Rathdowne Street, Carlton South, Victoria 3053, Australia
b
Prostate cancer has become the most common cancer of males in several developed countries. With increasing age, most men will develop microscopic foci of prostate cancer whether they live in a population at a high or low risk for the invasive form of the disease [1] and, although a majority of men will develop microscopic disease, only a small percentage of these slow-growing tumors will develop into invasive prostate cancer and an even smaller proportion will cause premature death. The principal focus of epidemiologic investigations of prostate cancer, therefore, has to be the identification of factors— amenable to intervention—that cause the common microscopic form to progress to invasive disease. Prostate cancer is highest in Western populations, particularly among the black population of the United States. The disease is uncommon in many Asian and populations of developing countries [2]. Consideration of the public health importance of prostate cancer should be tempered
This study was conducted within the framework of support from the Associazione Italiana per la Ricerca sul Cancro (Italian Association for Research on Cancer) (P. Boyle, G. Severi) and the National Health and Medical Research Council (940394), and was further supported by funding from Tattersall’s and The Whitten Foundation as well as infrastructure provided by The Cancer Council of Victoria (G.G. Giles). Many of the ideas presented herein have been published previously and elsewhere by the authors. Thus, it is important to note that the authors do not claim this as an original work but merely a compilation of ideas brought together for review purposes. * Corresponding author. E-mail address: [email protected]
with the observation that in many countries the average age at death from prostate cancer is approaching 80 years (Fig. 1). Indeed, as data from England and Wales demonstrate, the average age at diagnosis of prostate cancer is greater by a considerable margin than is the average age at diagnosis for other common cancers such as breast and colorectal. The epidemiology of prostate cancer has been notoriously difficult to study and the disease continues to present formidable challenges to epidemiologists. Much of the difficulty is linked with the lack of knowledge regarding disease specificity. Both the phenotypes and genotypes of prostate cancers are heterogeneous and studies that combine all forms of prostate cancer together are, therefore, likely to attenuate any associations that might only arise with particular subtypes. This problem has gained more widespread recognition in recent years, and epidemiologists have attempted to increase disease specificity in their studies, largely by stratifying on severity—for example, histologic grade, Gleason scores, stage of disease, progression, and death. Although this approach has occasionally produced strengthened associations, it has not greatly advanced etiologic understanding. The causes of prostate cancer have been investigated in numerous case-control studies and a few prospective cohort studies [3–5]. In addition to disease specificity, there have been other problems with epidemiologic studies of prostate cancer, particularly small sample sizes and, therefore, poor statistical power, poor exposure measurement, and inappropriate study designs. The best available epidemiologic evidence on prostate cancer can be obtained from only a handful of large well-controlled case-control
0094-0143/03/$ - see front matter Ó 2003, Elsevier Science (USA). All rights reserved. doi:10.1016/S0094-0143(02)00181-7
210
P. Boyle et al / Urol Clin N Am 30 (2003) 209–217
Fig. 1. Cumulative distribution of numbers of deaths from cancer in the United Kingdom, 1990.
studies and cohort studies. Although historically, case-control studies have identified numerous putative risk factors, only age and a family history of prostate cancer are well established. During the 1990s, prospective studies suggested that specific fatty acids, antioxidant vitamins, and carotenoids might alter prostate cancer risk. There also were reports that changes in plasma levels of key hormones and associated molecules, and naturally occurring variants in genes (polymorphisms) of the androgen, vitamin D, and insulin-like growth factor 1 (IGF-1) prostate cell growth regulatory pathways also might alter risk, and conjectures that dietary factors might modulate risk by interacting with these pathways. Population trends Prostate cancer is extremely age dependent, is uncommon before the age of 50, and increases rapidly thereafter. It demonstrates a large variation in incidence between countries. For example, the age-standardized incidence rates per 100,000 from the seventh edition of Cancer Incidence in Five Continents [2], which covers the period 1988–1992, ranged from very high in the United States (data derived from the Surveillance, Epidemiology and End Results (SEER) database of the National Cancer Institute: SEER blacks, 137; SEER whites, 101), to much lower in Europe (England and Wales, 28), to low in Asia (Japan, Miyagi Registry, 9). There are also substantial differences between ethnic groups living in the same country; for example, Japanese Americans have rates that are intermediate to those of the SEER whites and native Japanese stated above (Hawaii, 64; Los Angeles, 47) [2]. The changes in incidence observed in migrant populations are of interest because they point to factors that may increase prostate cancer risk and also to those that may reduce it.
Prostate cancer has increased in incidence in many countries around the world. In aging populations, the burden of prostate cancer can be expected to increase over time, even if the agespecific rates remain constant; however, between the mid 1980s and the early 1990s incidence rose dramatically, especially in parts of the United States [6]. The substantial changes in prostate cancer incidence and mortality that have occurred in the United States since the introduction of the prostate-specific antigen (PSA) blood test are of enormous interest and importance. Trends can be examined from 1975 [6]. Between 1986 and 1992, the overall age-adjusted incidence rate more than doubled from 86 to 179 per 100,000 per annum. Following this rise, incidence declined in all categories of age considered (under 65 years, 65– 74 years, and 75 years and older). The increase was most apparent in younger men (up to age 65), whereas the recent decline was most notable in men aged 75 and older. The increase in incidence between 1986 and 1992 occurred for both localized and regional stages of prostate cancer and was greatest in moderately differentiated tumors. The incidence of distant stage prostate cancer peaked in 1985 and has since declined [7]. From the pre-PSA era (1980–1985) until the PSA-era (1990–1995), the median age at diagnosis in the United States decreased by 1 year for whites and blacks. During the same period, the median age at death increased by 1 year in each racial group [7]. The median age at diagnosis in the United States (71 years), however, was considerably less than that in England and Wales (77 years) for the period 1988–1990 [8]. Undoubtedly, the trends of cancer incidence in the United States have been influenced by the widespread use of PSA testing. It may be premature, however, to attribute the recent reduction in prostate cancer mortality to PSA testing [9,10]. First, the reduction in mortality appeared to occur quickly to be confidently ascribed to screening and, second, the introduction of PSA was contemporaneous with that of widespread androgen deprivation therapy, a treatment that has had a significant impact on mortality. Family history and genetics Prostate cancer demonstrates a pattern of familial aggregation similar to that shown by breast and colorectal cancers, with an increased risk to a first-degree relative of a man with prostate cancer of about 2 to 3 on average [11–13],
P. Boyle et al / Urol Clin N Am 30 (2003) 209–217
the risk increasing the younger the age at diagnosis of the affected member. Even if a risk of 1.5 associated with having an affected first-degree relative was due to genetic factors, the combined effects of those genetic factors would have a large effect on disease risk, equivalent to an interquartile risk ratio an order of magnitude greater [14]. Thus, a similar degree of familial aggregation can be a consequence of either a rare high-risk genetic mutation or a common low-risk genetic polymorphism. Interest in the role of genetic factors in prostate cancer susceptibility has escalated due to a number of linkage analyses based on families that contain several men with prostate cancer, most of whom have early-onset disease. As discussed in a recent review [13], convincing replications have been rare, and heterogeneity analyses suggest that if any one of these regions contains a major prostate cancer gene, mutations in that gene would explain only a small proportion of multiple-case prostate cancer families. On the other hand, there have been reports of more modest risks of prostate cancer associated with common variants (polymorphisms) in candidate genes, such as those that encode the androgen receptor (AR), PSA, 5a-reductase type 2 (SRD5A2), cytochrome P450, vitamin D receptor (VDR), glutathione S-transferase, and HPC2/ELAC2 [15–21]. Thus far, early reports of associations have not been replicated in later studies—a pattern consistent with ‘‘data torture’’ of small clinical series. If true, however, the modest risks associated with common polymorphisms might explain a far greater proportion of disease than do the high risks associated with rare mutations. Hormones and other growth factors Growth and maintenance of normal prostate epithelium is regulated by the androgen and vitamin D pathways, with androgens stimulating and vitamin D metabolites inhibiting cell proliferation [22]. The androgen and vitamin D pathways interact at various levels, with one endpoint of both being the IGF-1 axis [22]. Perturbations of all three pathways have been associated with prostate cancer [22]. Cell division in the prostate is controlled by testosterone (T) that is reduced in the prostate to its active form, 5a-dihydrotestosterone (DHT), by 5-alpha reductase (SRD5A2) [23]. DHT binds to the AR to induce dimerization and binding to androgen response elements of
211
target genes to regulate their transcription [24]. Allelic variants of SRD5A2 (49T, 89V), which are thought to increase SRD5A2 activity have been associated with an increased risk of prostate cancer [25]. Short alleles of the AR CAG microsatellite also have been associated with increased risk and with cancers of aggressive phenotype [25]. AR genes with short CAG regions are more highly expressed compared with AR genes with longer CAG regions [26,27]. Other polymorphisms have been identified in genes encoding androgen biosynthetic and catabolic enzymes (such as CYP17, HSD3B2, and HSD17B3), but their association with prostate cancer has not been determined [26,28]. Consistent with the hypothesis that increased AR activity increases the risk of prostate cancer, prospective risk studies of androgen plasma/serum measurements suggest that a high plasma T to DHT ratio, high circulating levels of T, low levels of dehydroepiandrosterone or low levels of sex hormone binding globulin (SHBG)—which binds to T, decreasing its bioavailability—may elevate risk. Whereas androgens stimulate prostate cell proliferation, vitamin D inhibits it [29]. Two small nested case-control studies [30,31] have reported that high levels of the vitamin D metabolite 1,25D in prediagnostic sera are associated with lower risk of prostate cancer. Thus, alterations in the VDR gene that affect receptor activity might affect prostate cancer susceptibility. Three restriction fragment length polymorphisms as well as a polymorphism in the translation initiation site of the VDR gene, and a poly A length polymorphism have been identified [32,33]. It is not clear whether any of these affect VDR function, and although preliminary studies [33,34] implicate the poly A and the Taq I polymorphisms in prostate cancer risk, they need to be replicated. Both the AR and the VDR are thought to produce some of their growth effects via IGF-1. Binding proteins regulate IGF-1 availability in the prostate [22]. Expression of the major binding protein (IGF-BP3) is regulated by androgens and vitamin D; androgens inhibit expression of IGFBP3 [24], whereas an analogue of vitamin D has been shown to upregulate expression of IGF-BP3 [33]. The association between circulating IGF-1 levels and prostate cancer risk [3,4] was confirmed in a prospective study [4] that showed an approximate doubling of risk per 100-ng/mL increase in IGF-1, in serum samples collected before subsequent development of prostate cancer. This association strengthened when controlling for
212
P. Boyle et al / Urol Clin N Am 30 (2003) 209–217
IGF-BP3 [4]. Although a number of polymorphisms upstream of the IGF-1 transcriptional start site have been identified [34,35], direct association of these polymorphisms with prostate cancer risk has yet to be evaluated. Whites who are homozygous for the (CA) 19 alleles have significantly lower IGF-1 serum levels than do other genotypes [36]. Environmental factors Meat and fat The evidence of an association between meat and prostate cancer is more consistent than that for fat [37], particularly for advanced disease. The association with meat intake is difficult to separate from that of fat, however, because meat is a major vehicle for fat in Western diets. There is certainly insufficient evidence on which to base dietary advice. One possible explanation for the limited evidence is that meat produces carcinogenic heterocyclic amines when cooked at high temperature, and also produces polycyclic aromatic hydrocarbons when grilled over flames. The only study on this topic [38] has produced equivocal findings. Cohort studies [3,4] have reported positive associations between prostate cancer and red meat consumption, total animal fat consumption, and intake of fatty animal foods. It is thought that dietary fatty acids may affect risk in a variety of ways, including altering serum levels of sex hormones or modulating eicosanoid synthesis and affecting tumor cell proliferation, immune response, invasion, and metastasis; altering the composition of cell membrane phospholipids; affecting 5a-reductase type I activity [39]; forming free radicals from peroxidation; decreasing 1,25 D levels; and increasing IGF-1 levels. Evidence suggests that increased biosynthesis by prostate cancer cells of arachidonic acid-derived prostaglandins and hydroxyacid eicosanoids via the cycloxegenase type 2 (COX-2) enzyme pathway results in enhanced cancer cell proliferation and invasive and metastatic behavior. The COX enzyme may be inhibited directly by nonsteroidal anti-inflammatory drugs [40–42]. Plant foods, antioxidant vitamins, carotenoids, and minerals In general, vegetable consumption has not been strongly associated with prostate cancer risk [43], with the exception of a protective affect of tomatoes and foods that contain concentrated
tomato products. Lycopene, a fat-soluble carotenoid principally found in tomatoes, is an efficient singlet oxygen quencher, and has been shown to be unusually concentrated in the prostate gland [44]. In one cohort study [3,4], high lycopene intake was related to a 21% lower risk of prostate cancer (P<0.05), especially for advanced disease. In another study, the risk in men with higher plasma lycopene levels was reduced by 25%, and, for aggressive cancer, was reduced by 44%. Some studies support a possible protective effect of legumes, carrots, and green leafy and cruciferous vegetables, but this data is not compelling, nor is the evidence for other carotenoids. Vitamin A has been reported to both increase and reduce the risk of prostate cancer. Similarly, there is mixed evidence with regard to the effects of dietary beta-carotene: although some casecontrol studies have suggested a protective effect, no benefit was seen in large prospective studies. The evidence of a protective effect of vitamin E (a-tocopherol) supplements is more convincing. Male smokers randomized to take 50-mg a-tocopherol supplement in the Alpha-Tocopherol BetaCarotene (ATBC) Trial [45], had a statistically significant 32% decrease in clinical prostate cancer incidence, and a significant 41% reduction in prostate cancer mortality, compared with the placebo group. This should be interpreted with particular caution as this was not among the primary hypotheses for the study. Evidence of an effect associated with the amounts of a-tocopherol that can be obtained from dietary sources is weak. Another study [46] suggested that a-tocopherol is more potent and interacts with c-tocopherol. The consistent positive association between prostate cancer and dairy products that are rich in vitamin D may be explained by the high calcium content in dairy foods, which suppresses the formation and circulation of 1,25D levels. Two studies [4] have found strong positive associations between calcium intake and prostate cancer. Therefore, associations between fructose intake (negative) and meat intake (positive) and prostate cancer risk could be explained, in part, by effects on 1,25D levels. Phytoestrogens include two groups of hormonelike diphenolic compounds—isoflavonoids and lignans. Their consumption has been proposed as a contributing factor to the low levels of prostate and breast cancers seen in societies that consume high levels of soy products and other legumes. Phytoestrogens have been shown to act as antioestrogens, weak estrogens, and antioxi-
P. Boyle et al / Urol Clin N Am 30 (2003) 209–217
dants, and have been shown to inhibit prostate tumor model systems in vivo and in vitro [47]. Although their biologic function is not understood fully, they have been reported to inhibit 5a-reductase types 1 and 2 and 17b-hydroxysteroid dehydrogenase and aromatase, and to stimulate the synthesis of SHBG and uridine diphosphate glucuronyltransferase, suggesting that they may act, in part, by decreasing the biologically available fraction of androgens [47]. Very little analytical epidemiologic research is available [48]. Similarly, polyphenolic compounds in green tea, called catechins, are believed to have risk-reducing properties, but the epidemiologic evidence is scarce and inconsistent [49]. Selenium and zinc both have received attention as possible chemopreventive agents for prostate cancer, so much so that selenium (and vitamin E) supplementation is being tested in a randomized trial. The prostate has a unique relationship with zinc metabolism that is related to the zinc content of seminal fluid. A case-control study [50] has reported that zinc supplement use is associated in a significant dose–response fashion with a reduced risk of prostate cancer. Energy intake, energy balance, and body composition Energy intake may be associated with prostate cancer risk via an effect on IGF-1 levels [51]. Energy imbalance leads to adiposity. Two thirds of plasma E in men is converted from dehydroepiandrosterone and androstenedione by aromatase in adipose tissue and muscle. Thus, body composition affects the ratio of circulating E and T. Prostate cells are sensitive to T and E because they express both the a and b forms of the E receptor [52]. Continuous long-term exposure to a low level of E in combination with androgen during adulthood induces prostate enlargement in mice, rats, and dogs [53]. These findings suggest that an increased exposure to E at any time in life can affect prostate cell growth and prostate growth regulatory dysfunction. Associations between body mass index (BMI) and prostate cancer have been weak and inconsistent, perhaps reflecting the inadequacy of BMI as a measure of body composition. Physical activity theoretically could reduce prostate cancer risk by reducing plasma T levels. The evidence from epidemiologic studies is inconsistent, but suggestive of a protective effect, possibly restricted to high physical activity levels. For example, two recent prospective cohort
213
studies in the United States [51,52] showed no evidence of an association with physical activity, whereas another [54] showed that physically inactive men were at increased risk compared with very active men. This latter association, however, was limited to black Americans and was not statistically significant in whites. A cohort study of 22,895 Norwegian men gave a relative risk (RR) of 0.8 (confidence interval [CI]=0.62–1.03) for high versus low activity [55]. Physical activity is negatively correlated with obesity, and yet both are negatively correlated with T levels. There is only inconsistent evidence that obesity (as measured by BMI) is associated with prostate cancer. BMI is a problematic measure in this regard because it combines both adiposity and lean body mass—the two components having different hormonal associations, with the latter being under the influence of androgens and IGF-1. This issue will be resolved only by cohort studies that include separate estimates of lean and fat body mass [56]. Sexual factors Given that prostate growth and development is androgen dependent, several studies [5,57–60] have pursued the idea that prostate cancer risk might be related to variation in androgen milieu manifested by differences in sexual activity; however, these studies have reported only weak and inconsistent findings. One focus has been sexually transmitted diseases (STDs) and behaviors that are associated with increased risk of infection, such as multiple sex partners and sex with prostitutes. Apart from a history of STDs, there is limited evidence that papillomavirus infection increases risk [61]. The literature also is inconsistent with regard to sexual activity. Rotkin’s [62] proposal that reduced ejaculatory frequency in normal men increases the risk of prostate cancer is supported by a casecontrol study [63], which reported an odds ratio of 4.05 (CI=2.99–5.48) for men who had experienced an (undefined) period of interrupted sexual activity. Further support for Rotkin’s theory comes from a study that showed that presumably celibate Roman Catholic priests had an above-average risk of dying from prostate cancer [64]. Tobacco and alcohol Smoking is most likely not associated with prostate cancer incidence; however, there is some evidence that smoking may be positively associated with mortality from this cancer [65,66].
214
P. Boyle et al / Urol Clin N Am 30 (2003) 209–217
Recent cohort studies [67,68] gave estimates that were close to unity for associations between the incidence of all prostate cancer and either current or past smoking and modest positive associations cancer (RRs between 1.3 and 2) with respect to fatal prostate. A competent review [69] concluded that there was no association between low to moderate alcohol consumption and prostate cancer, but could not exclude the possibility of an association with heavy drinking or that the effects of alcohol might be observed in subgroups defined by genetic markers and family history. More recent findings [70–72] did not modify this conclusion. Occupation The accumulated studies of occupation have failed to provide compelling evidence of associations with the risk of prostate cancer. The weak associations in the literature are probably the result of poorly measured exposures and uncontrolled confounding with social class. Cadmium exposure once was highly suspected, but further follow-up of industrial cohorts failed to confirm evidence of an association. A number of studies have reported associations with farming—with particular reference to pesticides, but most have not measured exposures at the level of the individual. Summary The etiology of prostate cancer remains virtually unknown. Although there are a number of new leads with regard to risk factors for prostate cancer, more research is required to confirm them. There is little purpose in conducting further casecontrol studies of prostate cancer—particularly since the use of PSA testing has become widespread. Instead, future epidemiologic studies should focus on prostate tumor subclassification, in terms of method of detection, markers of biological ‘‘aggressiveness,’’ and genetic changes. Many of these new leads involve the possible influence of polymorphisms in key genes involved in important physiologic processes in the prostate. To fully explore the complexity of interrelationships between the several elements in these pathways will require large cohort studies in which blood is sampled prior to diagnosis. Such studies will be important for identifying which modifiable aspects of lifestyle (such as diet, alcohol, tobacco, physical activity) might be targeted for intervention, to reduce risk.
The detection of early prostate cancers by PSA testing relatives of men with prostate cancer has affected the prevalence of phenocopies and, hence, the meaningfulness of risk estimation in prostate cancer families. Because multiple-case families form the substrate for gene hunting via linkage analysis, this phenocopy phenomenon is going to cause considerable confusion and wasted effort. Presently, men with a family history of prostate cancer can be provided with little advice in terms of preventive action. It is likely that one or more genetic mutations associated with a high risk for prostate cancer will be identified in the near future. Even so, the risks probably will be similar to those for mutations in the first two breast cancer genes—informative for very few families. It is difficult to foresee, as and when high-risk mutation carriers are identified, what advice should be offered to them: prophylactic prostatectomies seem to have less attraction than do prophylactic mastectomies for women at high risk of breast cancer. This issue becomes more complex when considering counseling on the basis of a genetic profile involving many low-risk polymorphisms. Hopefully, such genetic screening should occur only after its efficacy has been established; when there is a better understanding of prostate biology, tumor heterogeneity, and prognosis; and when there are proven treatment or prevention options available. Prevention is held to be better than cure, and there are some potentially interesting chemopreventive agents that require careful trial before public health initiatives can be promoted. Potential candidates include vitamin E, selenium, zinc, and lycopene as dietary supplements. There are other agents that may be appropriate for pharmaceutical development, including inhibitors of COX-2 and IGF-1 activity. It is important that chemoprevention trials are followed-up for a sufficient period of time and that other endpoints also are captured, because the supplementation of diets with superphysiologic doses of individual micronutrients sometimes has caused unexpected and unwanted results—for example, an 18% increase in lung cancers observed in the beta-carotene arm of the ATBC trial [73].
References [1] Yatani R, Chigusa I, Akazaki K, et al. Geographic pathology of latent prostatic carcinoma. Int J Cancer 1982;29:611–6.
P. Boyle et al / Urol Clin N Am 30 (2003) 209–217 [2] Parkin DM, Whelan SL, Ferlay J, Raymond L, Young J, editors. Cancer incidence in five continents, Vol 7, Lyon, France; International Agency for Research on Cancer; 1997. [3] Clinton SK, Giovannucci E. Diet, nutrition, and prostate cancer. Annu Rev Nutr 1998;18:413–40. [4] Chan JM, Stampfer MJ, Giovannucci EL. What causes prostate cancer? A brief summary of the epidemiology. Semin Cancer Biol 1998;8:263–73. [5] Ross RK, Schottenfeld D. Prostate cancer. In: Schottenfeld D, Fraumeni JF Jr, editors. Cancer epidemiology and prevention. 2nd edition. New York: Oxford University Press; 1996. p. 1180–206. [6] Hsing AW, Devesa SS. Trends and patterns of prostate cancer: what do they suggest? Epidemiol Rev 2001;23(1):3–13. [7] Newcomer LM, Stanford JL, Blumenstein BA, et al. Temporal trends in rates of prostate cancer: declining incidence of advanced stage disease, 1974 to 1994. J Urol 1997;158:1127–30. [8] Boyle P, Severi G. Epidemiology of prostate cancer chemoprevention. Eur Urol 1999;35(5–6):370–6. [9] Tarone RE, Chu KC, Brawley O. Implications of stage-specific survival rates in assessing recent declines in prostate cancer mortality rates. Epidemiology 2000;11(2):167–70. [10] Oliver SE, Gunnell D, Donovan JL. Comparison of trends in prostate-cancer mortality in England and Wales and the USA. Lancet 2000;355(9217):1788–9. [11] Lesko SM, Rosenberg L, Shapiro S. Family history and prostate cancer risk. Am J Epidemiol 1996; 144:1041–7. [12] Hopper JL, Carlin JB. Familial aggregation of a disease consequent upon correlation between relatives in a risk factor measured on a continuous scale. Am J Epidemiol 1992;136:1138–47. [13] Ostrander EA, Stanford JL. Genetics of prostate cancer: too many loci, too few genes. Am J Hum Genet 2000;67:1367–75. [14] Peto J. Genetic predisposition to cancer. In: Cairns J, Lyon JL, Skolnick M, editors. Cancer incidence in defined populations (Banbury Report no. 4). Cold Spring Harbor (NY): Cold Spring Harbor Laboratory Press; 1980. p. 203–13. [15] Ingles SA, Ross RK, Yu MC, et al. Association of prostate cancer risk with genetic polymorphisms in vitamin D receptor and androgen receptor. J Natl Cancer Inst 1997;89:166–70. [16] Xue W, Irvine RA, Yu MC, et al. Susceptibility to prostate cancer: interaction between genotypes at the androgen receptor and prostate specific antigen loci. Cancer Res 2000;60:839–41. [17] Jaffe JM, Malkowicz SB, Walker AH, et al. Association of SRD5A2 genotype and pathological characteristics of prostate tumors. Cancer Res 2000; 60:1626–30. [18] Walker AH, Jaffe JM, Gunasegaram S, et al. Characterization of an allelic variant in the nifedipine-specific element of CYP3A4: ethnic distribu-
[19]
[20]
[21]
[22]
[23] [24] [25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
215
tion and implications for prostate cancer risk. Hum Mutat 1998;12:289. Ma J, Stampfer MJ, Gann PH, et al. Vitamin D receptor polymorphisms, circulating vitamin D metabolites, and risk of prostate cancer in US physicians. Cancer Epidemiol Biomarkers Prev 1998;242:467–73. Kelada SN, Kardia SL, Walker AH, et al. The glutathione S-transferase-mu and -theta genotypes in the etiology of prostate cancer: genotype– environment interactions with smoking. Cancer Epidemiol Biomarkers Prev 2000;9:1329–34. Rebbeck TR, Walker AH, Zeigler-Johnson C, et al. Association of HPC2/ELAC2 genotypes and prostate cancer. Am J Hum Genet 2000;67:1014–9. Russell PJ, Bennett S, Stricker P. Growth factor involvement in progression of prostate cancer. Clin Chem 1998;44:705–23. Santen RJ. Endocrine treatment of prostate cancer. J Clin Endocrinol Metab 1992;75:685–9. Bentel JM, Tilley WD. Androgen receptors in prostate cancer. J Endocrinol 1996;151:1–11. Ross RK, Pike MC, Coetzee GA, et al. Androgen metabolism and prostate cancer: establishing a model of genetic susceptibility. Cancer Res 1998; 58:4497–504. Choong CS, Kemppainen JA, Zhou Z-X, et al. Reduced androgen receptor gene expression with first exon CAG repeat expansion. Mol Endocrinol 1996;10:1527–35. Chamberlain NL, Driver ED, Miesfeld RL. The length and location of CAG trinucleotide repeats in the androgen receptor N-terminal domain affect transactivation function. Nucleic Acids Res 1994;22:3181–6. Moghrabi N, Hughes E, Dunaif A, et al. Deleterious missense mutations and silent polymorphism in the human 17b-hydroxysteroid dehydrogenase 3 gene (HSD17B3). J Clin Endocrinol Metab 1998; 83:2855–60. Issa LL, Leong GM, Eisman JA. Molecular mechanism of vitamin D receptor action. Inflamm Res 1998;47:451–75. Ingles SA, Ross RK, Yu MC, et al. Association of prostate cancer risk with genetic polymorphisms in vitamin D receptor and androgen receptor. J Natl Cancer Inst 1997;89:166–71. Taylor JA, Hirvonen A, Watson M, et al. Association of prostate cancer with vitamin D receptorgene polymorphism. Cancer Res 1996;56: 4108–4110. Corder EH, Friedman GD, Vogelman JH, et al. Seasonal variation in vitamin D, vitamin D-binding protein, and dehydroepiandrosterone: risk of prostate cancer in black and white men. Cancer Epidemiol Biol Prev 1995;4:655–9. Nickerson T, Huynh H. Vitamin D analogue EB1089-induced prostate regression is associated with increased gene expression of insulin-like
216
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42] [43]
[44]
[45]
[46]
[47]
[48]
[49] [50]
P. Boyle et al / Urol Clin N Am 30 (2003) 209–217 growth factor binding proteins. J Endocrinol 1999; 160:223–9. Miyao M, Hosoi T, Inoue S, et al. Polymorphism of insulin-like growth factor I gene and bone mineral density. Calcif Tissue Int 1998;63:306–11. Mullis PE, Patel MS, Brickell PM, et al. Constitutionally short stature: analysis of the insulin-like growth factor-I gene and the human growth hormone gene cluster. Pediatr Res 1991;29:412–5. Rosen CJ, Kurland ES, Vereault D, et al. Association between IGF I and a simple sequence repeat in the IGF I gene: implications for genetic studies of bone mineral density. J Clin Endocrinol Metab 1998;83:2286–90. Kolonel LN, Nomura AM, Cooney RV. Dietary fat and prostate cancer: current status. J Natl Cancer Inst 1999;91(5):414–28. Norrish AE, Ferguson LR, Knize MG, et al. Heterocyclic amine content of cooked meat and risk of prostate cancer. J Natl Cancer Inst 1999; 91:2038–44. Liang T, Liao S. Inhibition of steroid 5a-reductase by specific aliphatic unsaturated fatty acids. Biochem J 1992;285:557–62. Zhou JR, Blackburn GL. Bridging animal and human studies: what are the missing segments in dietary fat and prostate cancer? Am J Clin Nutr 1997;66:1572S–80S. Norrish AE, Jackson RT, McRae CU. Nonsteroidal anti-inflammatory drugs and prostate cancer progression. Int J Cancer 1998;77:511–5. Ghosii J, Myers C. Arachidonic acid metabolism and cancer of the prostate. Nutrition 1998;14:48–9. Chan JM, Giovannucci EL. Vegetables, fruits, associated micronutrients, and risk of prostate cancer. Epidemiol Rev 2001;23:82–6. Clinton SK, Emenhiser C, Schwartz SJ, et al. Cistrans lycopene isomers, carotenoids, and retinol in the human prostate cancer. Cancer Epidemiol Biomarkers Prev 1996;5:823–33. Rautalahti M, Albanes D, Virtamo J, et al. Betacarotene did not work: aftermath of the ATBC study. Cancer Lett 1997;114:235–6. Helzlsouer KJ, Huang HY, Alberg AJ, Hoffman S, Burke A, Norkus EP, et al. Association between alpha-tocopherol, gamma-tocopherol, selenium, and subsequent prostate cancer. J Natl Cancer Inst 2000;92(24):2018–23. Griffiths K, Denis L, Turkes A, Morton MS. Possible relationship between dietary factors and pathogenesis of prostate cancer. Int J Urol 1998; 5:195–213. Barnes S. Role of phytochemicals in prevention and treatment of prostate cancer. Epidemiol Rev 2001; 23(1):102–5. Chhabra SK, Yang CS. Tea and prostate cancer. Epidemiol Rev 2001;23:106–9. Kristal AR, Stanford JL, Cohen JH, et al. Vitamin and mineral supplement use is associated with
[51]
[52]
[53]
[54]
[55]
[56]
[57]
[58]
[59]
[60]
[61]
[62]
[63]
[64]
[65]
[66]
reduced risk of prostate cancer. Cancer Epidemiol Biomarkers Prev 1999;8:887–92. Putnam SD, Cerhan JR, Parker AS, et al. Lifestyle and anthropometric risk factors for prostate cancer in a cohort of Iowa men. Ann Epidemiol 2000; 10:361–9. Liu SE, Lee IM, Linson P, et al. A prospective study of physical activity and risk of prostate cancer in US physicians. J Epidemiol 2000;29:29–35. vom Saal FS, Timms BG, Montano MM, et al. Prostate enlargement in mice due to fetal exposure to low doses of estradiol or diethylstilbestrol and opposite effects at high doses. Proc Nat Acad Sci 1997;94:2056–61. Clarke G, Whittemore AS. Prostate cancer risk in relation to anthropometry and physical activity: the National Health and Nutrition Examination Survey Epidemiological Follow-Up Study. Cancer Epidemiol Biomarkers Prev 2000;9:875–81. Lund Nilsen TI, Johnsen R, Vatten LJ. Socioeconomic and lifestyle factors associated with the risk of prostate cancer. Br J Cancer 2000;82: 1358–1363. Severson RK, Grove JS, Nomura AMY, Stemmerman GN. Body mass and prostate cancer: a prospective study. BMJ 1988;297:713–5. Ewings P, Bowie C. A case-control study of cancer of the prostate in Somerset and east Devon. Br J Cancer 1996;74(4):661–6. Hayes RB, Pottern LM, Strickler H, et al. Sexual behaviour, STDs and risks for prostate cancer. Br J Cancer 2000;82(3):718–25. Andersson SO, Baron J, Bergstrom R, Lindgren C, Wolk A, Adami HO. Lifestyle factors and prostate cancer risk: a case-control study in Sweden. Cancer Epidemiol Biomarkers Prev 1996;5(7):509–13. Hsieh CC, Thanos A, Mitropoulos D, Deliveliotis C, Mantzoros CS, Trichopoulos D. Risk factors for prostate cancer: a case-control study in Greece. Int J Cancer 1999;80(5):699–703. Strickler HD, Goedert JJ. Sexual behavior and evidence for an infectious cause of prostate cancer. Epidemiol Rev 2001;23(1):144–51. Rotkin ID. Studies in the epidemiology of prostatic cancer: expanded sampling. Cancer Treat Rep 1977;61(2):173–80. Fincham SM, Hill GB, Hanson J, Wijayasinghe C. Epidemiology of prostatic cancer: a case-control study. Prostate 1990;17(3):189–206. Ross RK, Deapen DM, Casagrande JT, et al. A cohort study of mortality from cancer of the prostate in Catholic priests. Br J Cancer 1981; 43:233–5. Colditz G. Consensus conference: smoking and prostate cancer. Cancer Causes Control 1996;7: 560–2. Lumey LH. Prostate cancer and smoking: a review of case-control and cohort studies. Prostate 1996; 29:249–60.
P. Boyle et al / Urol Clin N Am 30 (2003) 209–217 [67] Giovannucci E, Rimm EB, Ascherio A, et al. Smoking and risk of total and fatal prostate cancer in United States Health professionals. Cancer Epidemiol Biomarkers Prev 1999;8:277–82. [68] Lotufo PA, Lee IM, Ajani UA, et al. Cigarette smoking and risk of prostate cancer in the Physicians’ Health Study. Int J Cancer 2000;87: 141–4. [69] Breslow RA, Weed DL. Review of epidemiologic studies of alcohol and prostate cancer: 1971–1996. Nutr Cancer 1998;30:1–13. [70] Lumey LH, Pittman B, Wynder EL. Alcohol use and prostate cancer in U.S. whites: no association in a confirmatory study. Prostate 1998;36:250–55.
217
[71] Villeneuve PJ, Johnson KC, Krieger N, Mao Y. Risk factors for prostate cancer: results from the Canadian National Enhanced Cancer Surveillance System. The Canadian Cancer Registries Epidemiology Research Group. Cancer Causes Control 1999;10:355–67. [72] Schuurman AG, Goldbohm RA, van den Brandt PA. A prospective study on consumption of alcoholic beverages in relation to prostate cancer incidence (The Netherlands). Cancer Causes Control 1999;10:597–605. [73] Rautalahti M, Albanes D, Virtamo J, et al. Betacarotene did not work: aftermath of the ATBC study. Cancer Lett 1997;114:235–6.