Downloaded from by guest on June 15, 2014

Epidemiologic Reviews
Copyright © 2001 by the Johns Hopkins University School of Hygiene and Public Health
All rights reserved
Vol. 23, No. 1
Printed in U.S.A.
Medical History and Etiology of Prostate Cancer
Edward Giovannucci 12
Clinical conditions and medical procedures may provide
insight as to the etiologies of some cancers. The clinical
course of prostate cancer is clearly influenced by hormonal
therapies; hormones may be involved in its etiology, and
various clinical conditions such as cirrhosis and diabetes
mellitus may be related to hormonal imbalances. A number
of studies have examined vasectomy in relation to risk of
prostate cancer, though no clear underlying mechanism has
emerged. Chronic inflammatory diseases cause cancer at a
variety of sites, but chronic prostatitis has not been studied
as a contributor to the risk of prostate cancer. Studies examining these conditions and procedures in relation to prostate
cancer risk will be briefly summarized here. Prostate cancers tend to be slow-growing and exhibit a wide range of
biologic potential with regards to aggressive behavior.
Thus, screening and diagnostic practices largely influence
the apparent incidence of prostate cancer. Because medical
conditions and procedures could impact on screening and
diagnostic tendencies, studies relating these factors to
prostate cancer risk should be interpreted with a great deal
of caution.
Cirrhosis may affect levels of a variety of hormones that
are produced or metabolized in the liver, including estrogens, testosterone, and insulin-like growth factors.
Cirrhosis induces changes in estrogen metabolism, and it is
known clinically that male cirrhotics may have hyperestrogenism severe enough to lead to testicular atrophy and
gynecomastia. Through such influences, cirrhosis may be
related to risk of prostate cancer. In 1964, Glantz (1) stud-
Received for publication August 31, 2000, and accepted for publication February 2, 2001.
Abbreviations: Cl, confidence interval; PSA, prostate-specific antigen; RR, relative risk.
Channing Laboratory, Department of Medicine, Brigham and
Women's Hospital and Harvard Medical School, Boston, MA.
Departments of Nutrition and Epidemiology, Harvard School of
Public Health, Boston, MA.
Correspondence to Dr. Edward Giovannucci, Channing
Laboratory, Harvard Medical School, 181 Longwood Avenue,
Boston, MA 02115 (e-mail: [email protected]
Downloaded from by guest on June 15, 2014
ied 350 autopsy cases with cirrhosis and 650 cases without
cirrhosis and found prostate cancer simultaneously present
in 3.3 percent of the former and 9.0 percent of the latter.
Glantz hypothesized that hyperestrogenism in cirrhosis
may have accounted for the lower prevalence of prostate
cancer in cirrhotics. In another autopsy study by Robson
(2), the prevalence of prostate cancer was somewhat
decreased in 205 cirrhotics compared with 2,227 noncirrhotics (8.3 percent versus 11.7 percent). No case of
prostate cancer was found in 49 instances of "severe" cirrhosis compared with 10 percent in the nonsevere group
consisting of 100 men (p < 0.01). However, there did not
appear to be an association between clinical signs of
hyperestrogenism and prevalence of prostate cancer,
though the numbers were small and blood levels of estrogen were not measured. Another autopsy study conducted
in Japan (3) also found that the prevalence of cirrhosis at
autopsy was significantly lower in prostate cancer cases
than in all male autopsy cases registered in Japan from
1958 to 1975. In 1,210 prostate cancer cases there were 37
cases of cirrhosis (3.1 percent), but there were 12,594
cases of cirrhosis out of 192,412 autopsy cases overall (6.5
Whereas the autopsy studies found an inverse relation
between cirrhosis and prostate cancer prevalence, this association was not supported in a follow-up study of 11,605
one-year survivors of cirrhosis identified from the Danish
National Registry of Patients from 1977 to 1989 (4).
Occurrence of any cancer through 1993 was determined by
linkage to the Danish Cancer Registry. For comparison, the
expected number of cancer cases was estimated from
national age-, sex-, and site-specific incidence cancer rates.
The standardized incidence ratio for prostate cancer was 1.0
(95 percent confidence interval (CI): 0.7, 1.3); there were 40
cases observed and 41.6 cases expected.
There are obvious limitations to autopsy studies; these are
based on prevalence measures of the two conditions at the
time of death. The relative timing of the occurrence of cirrhosis or prostate cancer cannot be known, making it impossible to infer causality. Also, the relation of autopsy-detected
to clinically-detected prostate cancer is complex and
unclear. Differences in the results among the studies may
also be related to the severity, duration, and cause of cirrhosis, which may all influence the endocrinologic disturbances. Future studies should account for these characteristics of cirrhosis and better document hormonal changes
related to cirrhosis not only of testosterone and estrogen, but
also of the insulin-like growth factors.
Diabetes mellitus
Vasectomy is used in many parts of the world as an effective method of contraception. This procedure is generally
regarded as safe and has had few documented acute or longterm adverse effects (13-15). However, in the recent past,
some studies have found that men who have had a vasectomy to be at moderately increased risk of prostate cancer
(16-25). For the most part, the elevations in risk among men
Another possibility for the frequent positive associations
is confounding by uncontrolled factors (34, 35). It is plausible that men with higher testosterone levels or who may
have a more "fertile" endocrine profile (36), who may be
inherently at higher risk of prostate cancer, may be more
likely to have a vasectomy. However, it is speculative that
significant enough, if any, differences in testosterone concentrations between men with and without vasectomy exist.
The biologic basis underlying the relation between vasectomy and prostate cancer remains speculative. Elevations in
anti-spermatozoa antibodies, decreased seminal hormone
concentrations, and decreased prostatic secretion have been
reported in men who have undergone vasectomy and in animal models (21). Whether these changes influence prostate
carcinogenesis remains unknown. A better understanding of
the effects of vasectomy on the prostate would help clarify
the biologic plausibility of the association. Future studies
should examine whether more consistent associations are
observed in specific subgroups of men who have had a
vasectomy. For example, some data suggest an increased
risk of prostate cancer limited to men who had their vasectomy at a relatively young age (20, 22, 25-27), or only after
several decades after the vasectomy was performed (21, 22,
26, 27). Consistent patterns in subgroups, though, have not
emerged thus far. Future studies should also take into
account screening patterns to examine the potential role of
detection bias.
Epidemiol Rev Vol. 23, No. 1, 2000
Downloaded from by guest on June 15, 2014
Several prospective studies have examined the relation
between type II diabetes mellitus with risk of prostate cancer and have found about a 30—70 percent reduction in
prostate cancer risk among diabetic men (5—7). The risk
appears to decrease over time since diagnosis of diabetes; in
male health professionals (7), a suggestive decrease in risk
occurred after 6 years postdiagnosis of diabetes, and a statistically significant decrease occurred only after 10 years
(relative risk (RR) = 0.5; 95 percent CI: 0.4, 0.8). Detection
bias did not appear to account for this association because
diabetics, despite their reduced risk, appeared to receive
even more screening for prostate cancer (7). However, no
clear evidence of an inverse association could be detected in
another prospective study (8) and in case-control studies (9,
10). Recently, an analysis of the Cancer Prevention Study of
the American Cancer Society (11) reported little association
between self-reported diabetes mellitus (types I and II) at
baseline and even a moderately elevated risk among men
who had had diabetes for 5 or more years.
The relation between diabetes and risk of prostate cancer
is likely to be quite complex. As a marker of hyperinsulinemia, diabetes should be associated with higher levels of circulating insulin and possibly higher free insulin-like growth
factor-1 because high insulin decreases one of the insulinlike growth factor-1 binding proteins. These factors would
presumably be growth-enhancing and may increase risk. On
the other hand, men with severe type II diabetes have substantially lower androgen levels, probably resulting from
toxic effects of hyperglycemia on the testosterone-producing
Leydig cells of the testis (12). Of note, hyperglycemia tends
to worsen and insulin levels tend to decrease over time, and
the risk of prostate cancer appears to decrease with increasing time since diagnosis of diabetes (7).
The complex timing of relations between type II diabetes
mellitus and specific hormonal changes may help explain
the discrepancies between the study of health professionals
(7) and the Cancer Prevention Study (11). The study of
health professionals found the inverse association primarily
among men who had had a diagnosis of diabetes mellitus for
at least 10-15 years. Also, that study was conducted in
1986-1994, a time period of considerably earlier detection
of prostate cancer due mostly to prostate specific antigen
(PSA) screening, compared with the Cancer Prevention
Study which took place from 1959 to 1972. It is important
for future studies to examine whether the longer time lag
between the diagnosis of diabetes and prostate cancer, and
the detection of prostate cancer at earlier stages, may
account for the apparent discrepancies.
who have had a vasectomy have been in the range of
1.5-2.0. This relation remains controversial because almost
an equal number of studies have found no association
(26-35). Overall, the epidemiologic literature is consistent
with either a weak positive association, or no appreciable
Given the weak nature of the association, bias and confounding need to be seriously considered. The possibility of
detection bias has not been adequately addressed. Most
studies were conducted prior to the widespread use of PSA
screening. Thus, differences in PSA screening between men
with and without vasectomy were unlikely to bias the
results. Nonetheless, the possibility remains that men who
had seen a urologist because of the vasectomy may have
been exposed to more digital rectal examinations subsequently, or were more likely to undergo invasive procedures required to diagnose prostate cancer. In a prospective
study conducted prior to PSA screening, enhanced detection by more frequent digital rectal examinations did not
appear to account for a positive association (21). In some
studies, vasectomy has been associated with advanced
stage of disease (18, 21, 22, 25) or symptomatic cancers
(16) arguing against detection bias, which would typically
produce stronger associations with less advanced disease.
However, two recent studies conducted largely in the "PSA
era" have reported suggestive associations only with
earlier-stage, lower-grade prostate cancer (26, 27). Positive
associations have also been reported for non-Western populations (18, 25), for which screening should not have been
an issue. Unfortunately, no study has reported on risk of
prostate cancer mortality, which would be less prone to
detection bias.
Medical History and Prostate Cancer Etiology
Chronic inflammation of the prostate
Epidemiol Rev Vol. 23, No. 1, 2000
epidemiologic studies. Until a valid practical definition can
be used for population-based studies, this hypothesis will be
difficult to study from an epidemiologic perspective. An
inflammatory etiology of prostate cancer is important to
establish because this could enhance preventive efforts, particularly if a specific agent is identified. Although a cause of
inflammation is not yet apparent, it is critical to note that
anti-inflammatory and antioxidant mechanisms are likely to
play an important anticarcinogenic role if chronic inflammation is important. An inflammatory process could possibly account for the suggestive evidence that antioxidants
(e.g., lycopene, vitamin E, selenium) and that nonsteroidal
anti-inflammatory drugs such as aspirin may be protective
Prostate cancer is a hormonally-related cancer; thus, it is
quite plausible that clinical syndromes that influence hormonal levels also affect risk of prostate cancer. There has
been a modest degree of study related to diabetes mellitus
and cirrhosis, and while the data are suggestive of a protective association for each, the studies are inconsistent. The
complex natural history of such diseases and their varying
impact on hormones complicates the interpretation of results
from such studies. Investigators who have access to existing
databases which have substantially large numbers to examine these associations are encouraged to evaluate these
hypotheses. However, prospective studies with archived
blood samples which can directly examine hormonal
hypotheses are preferable. A relation between vasectomy
and prostate cancer cannot be ruled out at this time, but any
association is likely to be relatively modest. Nonetheless, as
vasectomy remains an important contraceptive option,
further study of this relation, particularly in potentially susceptible subgroups, is warranted. There are almost no epidemiologic data on chronic prostatitis and risk of prostate
cancer. As it is quite plausible that prostatitis is carcinogenic, as evidenced by carcinogenic effects of chronic
inflammation in other organs, study of prostatitis in relation
to prostate cancer should be undertaken.
1. Glantz GM. Cirrhosis and carcinoma of the prostate gland. J
Urol 1964,91:291-3.
2. Robson MC. Cirrhosis and prostatic neoplasms. Geriatrics
3. Araki H, Mishina T, Miyakoda K, et al. An epidemiological
survey of prostatic cancer from the Annual of the
Pathological Autopsy Cases in Japan. Tohoku J Exp Med
1980; 130:159-64.
4. S0rensen HT, Friis S, Osen JH, et al. Risk of liver and other
types of cancer in patients with cirrhosis: a nationwide cohort
study in Denmark. Hepatology 1998;28:921-5.
5. Thompson MM, Garland C, Barrett-Connor E, et al. Heart
disease risk factors, diabetes, and prostatic cancer in an adult
community. Am J Epidemiol 1989; 129:511-17.
6. Adami HO, McLaughlin J, Ekbom A, et al. Cancer risk in
patients with diabetes mellitus. Cancer Causes Control
Downloaded from by guest on June 15, 2014
Long-standing chronic inflammation by diverse agents is
a cause of carcinoma of the liver, large bowel, urinary bladder, esophagus, and stomach. Chronic inflammatory states
are characterized by activated phagocytic macrophages,
which produce bactericidal reactive oxygen and nitrogen
species. Inflammatory states cause substantial potential for
mutational and promoting effects through exposure to
highly reactive compounds and growth factors, and
increased cell turnover. Given the strong link between
inflammatory states and carcinogenesis, prostatitis is an
obvious candidate to consider with regards to carcinogenic
potential, although there has been essentially no epidemiologic study of this relation.
Prostatitis is an inflammation or infection of the prostate
typically characterized by pain and discomfort in the rectal
or perineal region. This condition, common in adult men, is
categorized as acute bacterial, chronic bacterial, and nonbacterial. One community-based population (37) estimated
a prevalence of physician-diagnosed prostatitis of 11 percent
based on medical record documentation, which may be
incomplete. The actual prevalence of prostatitis, including
undiagnosed cases, is likely to be much higher. In acute and
chronic bacterial prostatitis, the causative agents are generally gram-negative bacteria, usually Escherichia coli, that
gain access to the prostate by means of urinary reflux during
a urinary tract infection. In nonbacterial prostatitis, inflammatory cells are present but a causative bacterial agent cannot be cultured in prostatic secretions or urine. Nonbacterial
prostatitis, often called idiopathic prostatitis, is by far the
most common form of prostatitis, especially in older men.
Novel infectious agents may be involved, as detected by the
isolation and sequencing of prokaryotic ribosomal DNA in
idiopathic prostatitis (38).
Recent pathologic evidence summarized by De Marzo et
al. (39) suggests that chronic inflammation may contribute
to prostate carcinogenesis. Focal prostatic glandular atrophy, which has been suggested as a potential precursor to
prostatic adenocarcinoma, is closely associated with chronic
inflammation. De Marzo et al. have shown that almost all
forms of focal atrophy are proliferative in nature, and,
hence, termed proliferative inflammatory atrophy. The vast
majority of proliferative inflammatory atrophy lesions are
associated with inflammation, consisting of an infiltrate of
lymphocytes and macrophages. De Marzo et al. postulate
that this common lesion arises in the setting of increased
oxidative stress, likely derived from proximate inflammatory cells. Many of the proliferating cells appear to have an
immature secretory cell phenotype, a phenotype similar to
prostatic intraepithelial neoplasia and prostate cancer. In
addition, there are elevated levels of glutathione-5transferase PI expression, suggestive of increased electrophile or oxidative stress.
The epidemiologic literature regarding prostatitis and risk
of prostate cancer is extremely sparse. There are no reports
from case-control or cohort studies that have examined
well-documented chronic bacterial or idiopathic prostatitis
in relation to prostate cancer risk. Perhaps this lack of study
relates to the difficulty of characterizing the exposure for
risk of prostate cancer. Am J Epidemiol 1990;132:1051-5.
24. Spitz MR, Fueger JJ, Babaian RJ, et al. Re: "Vasectomy and
risk of prostate cancer." (Letter). Am J Epidemiol 1991;
25. Platz EA, Yeole BB, Cho E, et al. Vasectomy and prostate
cancer: a case-control study in India. Int J Epidemiol 1997;
26. Lesko SM, Louik C, Vezina R, et al. Vasectomy and prostate
cancer. J Urol 1999;161:1848-52.
27. Stanford JL, Wicklund KG, McKnight B, et al. Vasectomy
and risk of prostate cancer. Cancer Epidemiol Biomarkers
Prev 1999;8:881-6.
28. Rosenberg L, Palmer J, Zauber A, et al. The relation of
vasectomy to the risk of cancer. Am J Epidemiol 1994; 140:
29. John EM, Whittemore AS, Wu AH, et al. Vasectomy and
prostate cancer: results from a multiethnic case-control study.
J Natl Cancer Inst 1995;87:662-9.
30. Sidney S, Quesenberry CP, Sadler MC, et al. Vasectomy and
the risk of prostate cancer in a cohort of multiphasic healthcheckup examinees: second report. Cancer Causes Control
31. Nienhuis H, Goldacre M, Seagroatt V, et al. Incidence of disease after vasectomy: a record linkage retrospective cohort
study. BMJ 1992;304:743-6.
32. Moller H, Knudsen LB, Lynge E. Risk of testicular cancer
after vasectomy: cohort study of over 73,000 men. BMJ
33. Zhu K, Stanford JL, Daling JR, et al. Vasectomy and prostate
cancer: a case-controlled study in a health maintenance organization. Am J Epidemiol 1996;144:717-22.
34. Skegg D. Vasectomy and prostate cancer: is there a link? N
35. Howards SS, Peterson HB. Vasectomy and prostate cancer:
chance, bias, or a causal relationship. JAMA 1993;269:
36. Armenian HK, Lilienfeld AM, Diamond EL, et al.
Epidemiologic characteristics of patients with prostatic neoplasms. Am J Epidemiol 1975;102:47-54.
37. Roberts RO, Lieber MM, Rhodes T, et al. Prevalence of a
physician-assigned diagnosis of prostatitis: The Olmsted
County Study of Urinary Symptoms and Health Status
among Men. Urology 1998;51:578-84.
38. Riley DE, Berger RE, Miner DC, et al. Diverse and related
16S rRNA-encoding DNA sequences in prostate tissues of
men with chronic prostatitis. J Clin Microbiol 1998;36:
39. De Marzo AM, Marchi VL, Epstein JI, et al. Proliferative
inflammatory atrophy of the prostate: implications for prostatic carcinogenesis. Am J Pathol 1999;155:1985-92.
40. Norrish AE, Jackson RT, McRae CU. Non-steroidal antiinflammatory drugs and prostate cancer progression. Int J
Cancer 1998;77:511-15.
Epidemiol Rev Vol. 23, No. 1, 2000
Downloaded from by guest on June 15, 2014
7. Giovannucci E, Rimm EB, Stampfer MJ, et al. Diabetes mellitus and risk of prostate cancer (United States). Cancer
Causes Control 1998;9:3-9.
8. Coughlin SS, Neaton JD, Sengupta A. Cigarette smoking as
a predictor of death from prostate cancer in 348,874 men
screened for the Multiple Risk Factor Intervention Trial. Am
J Epidemiol 1996;143:1002-6.
9. Mishina T, Watanabe H, Araki H, et al. Epidemiological
study of prostate cancer by matched-pair analysis. Prostate
10. La Vecchia C, Negri E, Franceschi S, et al. A case-control
study of diabetes mellitus and cancer risk. Br J Cancer 1994;
11. Will JC, Vinicor F, Calle EE. Is diabetes mellitus associated
with prostate cancer incidence and survival? Epidemiology
12. Ando S, Rubens R, Rottiers R. Androgen plasma levels in
male diabetics. J Endocrinol Invest 1984;7:21-4.
13. World Health Organization. Meeting to evaluate research
needs and priorities regarding the relationship of vasectomy to
cancers of the testis and prostate. Special Programme of
Research, Development and Research Training in Human
Reproduction. Geneva, Switzerland: World Health
Organization, 1991.
14. Giovannucci E, Tosteson TD, Speizer FE, et al. A long-term
study of mortality in men who have undergone vasectomy. N
Engl J Med 1992;326:1392-8.
15. Massey FJ Jr., Bernstein GS, O'Fallon WM, et al. Vasectomy
and health: results from a large cohort study. JAMA 1984;
16. Lightfoot N, Krieger N, Sass-Kortsak A, et al. Prostate cancer risk: medical history, sexual, and hormonal factors.
(Abstract). Ann Epidemiol 2000;10:470.
17. Honda GD, Bernstein L, Ross RK, et al. Vasectomy, cigarette
smoking, and age at first sexual intercourse as risk factors for
prostate cancer in middle-aged men. Br J Cancer 1988;57:
18. Hsing AW, Wang RT, Gu FL, et al. Vasectomy and prostate
cancer risk in China. Cancer Epidemiol Biomarkers Prev
19. Mettlin C, Natarajan N, Huben R. Vasectomy and prostate
cancer risk. Am J Epidemiol 1990;132:1056-61.
20. Hayes RB, Pottern LM, Greenberg R, et al. Vasectomy and
prostate cancer in US blacks and whites. Am J Epidemiol
21. Giovannucci E, Tosteson TD, Speizer FE, et al. A retrospective cohort study of vasectomy and prostate cancer in US
men. JAMA 1993;269:878-82.
22. Giovannucci E, Ascherio A, Rimm EB, et al. A prospective
cohort study of vasectomy and prostate cancer in US men.
JAMA 1993;269:873-7.
23. Rosenberg L, Palmer JR, Zauber AG, et al. Vasectomy and the