Prostate cancer is the most commonly diagnosed cancer and the second leading cause of cancer death among men in the United States.1 The introduction of screening by prostate-specific antigen (PSA) and the considerable media attention directed toward prostate cancer in the late 1980s and early 1990s contributed to a tremendous increase in the incidence of diagnosed prostate cancer. Incidence rates have stabilized somewhat in recent years, but still remain higher than rates in the pre-PSA era.2,3
Despite the large public health impact of prostate cancer, very little is known about risk factors for the condition. Established risk factors include age, race or ethnicity, and family history of prostate cancer. Several other risk factors have been suggested,4 including the consumption of red meat, dairy products, total fat, and high animal fat.5–7 Selenium,8–10 vitamin E,11,12 lycopene,13 carotene, and green vegetables could be protective. However, results for these dietary factors are inconsistent.14–17 With the increase in the proportion of men living to age 80 years and older, the impact of prostate cancer will continue to increase, and the need to identify preventable risk factors for the condition becomes more urgent.
Several investigators have hypothesized that, in addition to other risk factors, prostate cancer may have an infectious etiology. This hypothesis stems from observations of a positive association between prostate cancer and viral infections,18–22 sexually transmitted disease,23–25 number of sexual partners17,23,26 and frequency of sexual encounters with prostitutes.24,25 These findings are not consistent and other studies have found no associations with some of these factors24,26–35 Yet, regular use of condoms has been found to be protective.25 It is unclear whether a possible association between prostate cancer and infection is the result of an infectious agent directly or a continuing inflammation associated with the infection.
Prostatitis can be associated with an infection (acute or chronic bacterial prostatitis) or inflammation with no demonstrable infection (chronic pelvic pain syndrome inflammatory type). In some cases, there could be chronic pelvic pain (noninflammatory type) of uncertain relation to the prostate.36 Despite the infectious etiology hypothesis for prostate cancer, the association between prostate cancer and prostatitis has not been thoroughly examined. Although a positive association between prostatitis and prostate cancer has been reported in some studies37–41 and in a meta-analysis assessing the association,42 other studies found no association.41,43–45 The inconsistency in the findings could, in part, relate to choice of control subjects, multiple comparisons, or the definition of prostatitis. The purpose of this study was to conduct a population-based case-control study to investigate the association between prostate cancer and prostatitis (infection or inflammation of the prostate).
The study was conducted in Olmsted County, Minnesota, where population-based epidemiologic research has been conducted for many years. Several unique factors enhance the feasibility of such studies. As a result of the relative isolation of the community from other urban centers, along with the availability of high-quality specialty care at the Mayo Clinic, most residents seek care from a small number of providers within the county, primarily the Mayo Clinic and the Olmsted Medical Center. The potential for epidemiologic research is further enhanced by the use of unit medical records through which all outpatient and inpatient information (clinic, emergency room, hospital) for patients is contained in the same patient folder. This has permitted collection of comprehensive and detailed clinical information over extended periods of time for county residents. These medical records can be easily retrieved using patient identifiers and an extensive indexing system based on surgical and medical diagnoses, which is maintained by the Mayo Clinic. Through the Rochester Epidemiology Project,46,47 a similar indexing system was created for the other providers of care in Olmsted County. This project permits the linkage of all medical record information from all sources of medical care available in the community and used by the Olmsted County population. Consistent with the Mayo Foundation Institutional Review Board requirements, no medical records were reviewed for men who had refused to provide authorization for the use of their medical records in research. The study protocol was reviewed and approved by the Mayo Foundation Institutional Review Board.
Selection of Cases
Olmsted County residents with new, histologically confirmed diagnoses of prostate cancer between January 1, 1980, and December 31, 1996, were identified using the medical and surgical index linkages of the Rochester Epidemiology Project. We also searched the Mayo Clinic Tumor Registry to identify any additional cases of prostate cancer not identified from the medical and surgical index. In addition, the surgical index was searched to identify men who had a radical prostatectomy during the study period. For each potential case, we reviewed the complete medical record to confirm the diagnosis of prostate cancer and to establish the date of diagnosis of prostate cancer (index date). We identified 842 men from whom a 50% stratified random sample was selected. Subjects were stratified on year of diagnosis within 3 time periods, resulting in 108 cases from 1980 to 1987, 192 from 1988 to 1992, and 121 from 1993 to 1996, for a total of 421 cases.
Selection of Control Subjects
A set of control subjects was identified from the Mayo Clinic databases using the Rochester Epidemiology Project to identify all male residents of Olmsted County who were born in 1955 or earlier (ie, age between 23 and 100 years) and who had received medical care in the county between 1979 and 1997. We selected 2 control subjects for each prostate cancer case subject, matching by year of birth, duration of the medical record in the community, registration for medical care within 1 year of the index date, and residency in Olmsted County on the index date.
We excluded cases (and their matched control subjects) if the case did not provide authorization for the use of his medical records for research, was not a resident in Olmsted County, or was being held in the local federal prison at the date of diagnosis of prostate cancer. Thus, from the 421 eligible cases, we included 409 cases and 803 control subjects in the study, 394 prostate cancer cases with 2 matched control subjects each and 15 prostate cancer cases with a single matched control subject.
We reviewed all community medical records for cases and control subjects to determine a history of prostatitis. For each episode, medical record information was obtained on the general constitutional and specific urogenital symptoms presenting at that visit. Acute symptoms assessed included fever or chills, myalgia, dysuria, frequency and urgency of urination, and hematuria. Chronic pain symptoms assessed were suprapubic, lower abdominal or pelvic pain, perineal pain or discomfort, lower back pain, penile pain, scrotal or testicular pain, rectal or perirectal pain, and painful ejaculation or postejaculatory pain. Acute urogenital infections assessed from the physician notes were urethritis, cystitis, trigonitis, epididymitis, orchitis, or urinary tract infection. A key sign was tenderness on digital rectal examination. We abstracted laboratory records for each episode to assess evidence of urinary or prostatic infection based on results of urinalysis, and culture, microscopy, and Gram stain of urine and of expressed prostatic secretions. Laboratory evidence of infection was a positive Gram stain or urinalysis (>100,000 bacteria/mL) and a positive culture for bacteria in the urine or prostatic secretions with or without white blood cells (>10 cells/high-power field or >1000/mL). A history of recurrent urinary tract infection (UTI), defined as 2 or more laboratory-confirmed episodes of UTI in a given year, was also assessed from laboratory test results and physician notes. Any hospitalization or intravenous therapy for treatment of a prostatitis episode was noted. The dates and frequency of serum PSA determinations were also abstracted.
Characterization of Prostatitis
We characterized an episode as acute prostatitis, chronic bacterial prostatitis, chronic nonbacterial prostatitis, or questionable prostatitis based on prespecified criteria. Acute prostatitis was defined by the presence of a tender digital rectal examination or presence of any acute symptoms, together with any of the following: laboratory evidence of a urogenital infection, hospitalization or intravenous therapy for treatment of an acute episode of prostatitis, or physician notation of a urogenital infection (see previously). Chronic bacterial prostatitis was characterized by a history of recurrent UTI or chronic pain symptoms and laboratory evidence of urinary or prostatic infection as defined previously. An episode was characterized as chronic nonbacterial prostatitis (specifically, chronic pelvic pain syndrome inflammatory type) if it did not meet the criteria for acute or chronic bacterial prostatitis, and there was a history of chronic symptoms and laboratory evidence of inflammation without laboratory evidence of infection.
The intrarater reliability of data was assessed for 50 study subjects. Medical records for these subjects were reabstracted 3 to 6 weeks after completion of the initial abstraction by the same nurse abstractor using the same methods. The percent agreement was between 91% and 98% for each of the main response measures (acute bacterial prostatitis, chronic bacterial prostatitis, and chronic pelvic pain syndrome).
Using conditional logistic regression methods, taking into account the case-control matching, we investigated the associations of prostate cancer with a history of acute bacterial prostatitis, chronic bacterial prostatitis, chronic pelvic pain syndrome (inflammatory type), and any prostatitis. We explored the potential impact of age at index date as an effect modifier by including age-quartile-by-prostatitis interaction terms in the conditional logistic regression analyses, and in similar analyses stratified by index age quartiles. In addition, we assessed whether the association between prostate cancer and prostatitis varied with the number of episodes of prostatitis or the time from first episode of prostatitis to index date (2–10 years and greater than 10 years) using logistic regression analyses. Because both prostatitis and prostate cancer are associated with an elevated serum PSA level, the initiation of PSA screening in 1987 could contribute to increased prostate cancer detection in men with prostatitis, thus confounding the association. Therefore, we conducted stratified analyses by time period (index date before 1987 and index date in 1987 or later) and by number of PSA determinations up to 1 year before the index date (0 vs. ≥1) to test for their potential influence on the association of interest. In a further attempt to evaluate potential surveillance bias, we tested for a difference in the mean time interval from the last episode of prostatitis to the index date in cases diagnosed before 1987 and those diagnosed in 1987 or later; we also repeated all the analyses with exclusion of episodes of prostatitis less than 2 years before the index date. These episodes could be associated with increased surveillance and potential detection bias for prostate cancer.
The characteristics of cases and control subjects are presented in Table 1. The median age was 70 at index date in both groups, with similar median duration of the medical record. There were 174 unique episodes of prostatitis noted in the medical records of study subjects; 47 (27%) were categorized as acute, 37 (21%) as chronic bacterial, 58 (33%) as chronic without evidence of a bacterial infection (chronic pelvic pain syndrome inflammatory type), and 32 (18%) could not be categorized. The 142 classifiable episodes occurred in 91 study subjects (41 cases and 50 control subjects). Sixty-two men had a single episode (24 acute, 21 chronic bacterial, and 17 chronic nonbacterial) and 29 men had multiple episodes. The mean time from a first episode of prostatitis to the index date was 16.4 (standard deviation [SD], 14.3 years) in the 91 men. The median age at first episode of any type of prostatitis was 53.1 (25th percentile 38.8, 75th percentile 64.4) years overall for both cases and control subjects and did not differ substantially between case and control subjects (54.6 and 51.8 years, respectively). The median age at first episode of acute prostatitis (59.0 years) was higher than that for chronic bacterial (50.1 year) or chronic nonbacterial (50.3) prostatitis.
A diagnosis of prostate cancer was most strongly associated with a history of acute prostatitis (Table 2). The relative odds of prostate cancer in men who had ever had an episode of acute prostatitis was 2.5 (95% confidence interval [CI] = 1.3–4.7). The relative odds of prostate cancer in men with a history of chronic bacterial prostatitis was elevated but not as strong (1.6; 0.8–3.1). There was no association with chronic nonbacterial prostatitis (0.9; 0.4–1.8).
The relative odds of prostate cancer in men with a history of any type of prostatitis was 1.7 (1.1–2.6). When men with questionable episodes were included in the model, the estimates associated with any type of prostatitis declined slightly (odds ratio [OR] = 1.6; CI = 1.1–2.5). This association was further reduced after exclusion of episodes that occurred less than 2 years from the index date (1.5; 0.9–2.3). The exclusion of recent episodes also reduced the association with acute prostatitis to 1.9 (CI = 0.9–3.8). In analyses stratified by index age quartiles, the relative odds of prostate cancer in men with a history of acute or of chronic bacterial prostatitis was highest in men aged 70 to 77 years at diagnosis with very wide confidence intervals. The relative odds of acute prostatitis were 1.3 (CI = 0.4–4.7) for men age 39 to 63 years, 2.8 (0.8–10.0) for age 64 to 70, 5.9 (1.2–28.8) for age 70 to 77, and 1.8 (0.5–7.4) for age 78 and older, and 1.4 (0.4–4.7), 1.1 (0.6–57.7), 6.0 (0.6, 57.7), and 0.5 (0.1–5.4), respectively, for chronic bacterial prostatitis.
The association between prostate cancer and prostatitis was strongest after a single episode of prostatitis. The relative odds of prostate cancer in men with a single episode only were 2.3 (CI = 1.4–3.8) and in men with 2 or more episodes were 0.8 (0.3–1.8), compared with men who had no history of prostatitis. The odds ratio for men with a first diagnosis of any type of prostatitis 2 to 10 years from the index date was 2.0 (0.98–4.10) and for greater than 10 years from the index date was 1.2 (0.6–2.1), compared with men who had no history of prostatitis (excluding episodes within 2 years of the index date).
The association between acute prostatitis and prostate cancer persisted after excluding recent episodes of prostatitis. There was no effect attributable to increased PSA determinations in recent years. The mean time from last episode of acute prostatitis to date of diagnosis of prostate cancer was 13.7 (SD = 11.6) years in prostate cancer cases diagnosed before 1987 and 11.6 (SD = 11.9) years in cases diagnosed in 1987 or later. This duration increased to 16.5 and 14.0 years, respectively, when episodes less than 2 years before the index date were excluded. Adjustment for number of serum PSA determinations did not alter the association between prostate cancer and any of the types of prostatitis.
In analyses stratified by time period, there was no significant interaction for any of the prostatitis categories, although the power for such effects is likely limited as a result of small sample sizes. The overall association of prostatic cancer with any prostatitis was similar in both time periods (Table 3). The association with acute prostatitis, however, appeared slightly stronger in the more recent time period (OR = 3.3) than in the years before 1987 (OR = 1.1). When the episodes in the 2 years before the index date were excluded, the odds ratios for any prostatitis and for acute prostatitis were slightly attenuated. In contrast, the odds ratios for chronic episodes were slightly lower in the more recent time period.
These data do not provide compelling evidence for a role of chronic inflammation with prostate cancer, because the estimates for chronic bacterial prostatitis were only slightly elevated. These findings cannot conclusively refute the hypothesis of an infectious etiology for prostate cancer, however. The lack of an association with chronic prostatitis could be the result of inadequate power to detect a significant difference or chronic prostatitis being a less precise diagnosis compared with acute prostatitis.
The association with acute prostatitis could be the result of more frequent contact with the medical system. Acute prostatitis necessitates a visit to the doctor, which could increase the likelihood of a diagnosis of prostate cancer. We examined this potential for detection bias by excluding all episodes of acute prostatitis that occurred less than 2 years before the diagnosis of prostate cancer. Although the odds ratio remained elevated, the association was not as strong and the confidence interval included 1.0. Acute prostatitis is also associated with an elevation in serum PSA level and therefore could introduce detection bias in the PSA screening era (1987 and later). Stratifying the analysis by time period showed a slightly stronger association in the more recent time period, which could suggest a potential effect of detection bias. However, the long lag time between the last episode of acute prostatitis and the index date (mean of 12 years), as well as the absence of confounding as a result of the number of PSA determinations, suggests that the findings were not primarily the result of bias from frequent contact with the medical system. Alternatively, the association with acute prostatitis could be the result of other factors such as an altered immune mechanism that increases the risk of both acute prostatic infection and prostate cancer.
With an underlying hypothesis that chronic inflammation could be associated with cancer, we expected to see a dose-response relation with measures of chronicity. When chronicity was assessed as multiple episodes or “chronic prostatitis,” there was no evidence of a dose-response association. However, with only 29 men who had multiple episodes of prostatitis in our study, the power to detect a dose-response association was limited.
Several other studies support the potential role of an infectious agent in the etiology of prostate cancer. Moyret-Lalle et al.18 reported a high prevalence of detectable HPV-16 DNA in prostate tumors. Dillner et al. observed a 2.6-fold increased risk of prostate cancer in men with antibodies to HPV-18.20 Other studies, however, have shown no association between HPV DNA or HPV-16 seroprevalence and prostate cancer.27–29,33 The absence of an association in these studies could in part be the result of the relatively small sample sizes. Sexually transmitted disease has also been associated with prostate cancer.4,17,23–25 Men with a history of gonorrhea or syphilis had a 60% increased risk of prostate cancer.25 Other suggested risk factors include higher lifetime number of sexual partners17,23,26 and frequent sexual encounters with prostitutes.25 Prokaryotic DNA sequences have been identified in the prostatic tissue of men with prostatitis.48,49 Although these organisms could be involved in the pathogenesis of prostate cancer, their precise role is unclear.50
The protective effect of the regular use of condoms on the risk of prostate cancer and a recently published metaanalysis, which estimated a 60% increased risk of prostate cancer in men with a history of prostatitis,42 further support the potential role of an infectious etiology.25 Several of the studies that have examined the association between prostatitis and prostate cancer have been population-based38,41,43–45 or hospital-based40,51–53 case-control studies. Positive associations were noted in some studies,38–41 no associations in some,43–45,52 and an inverse association in others.51,53 In a population-based case-control study, Rosenblatt et al. reported a nonsignificant but elevated risk of prostate cancer in men with prostatitis diagnosed less than 2 years before the index date.23 No association was observed with earlier diagnoses of prostatitis (>2 years), but it is not clear how prostatitis was assessed in that study. The lack of definitive diagnostic criteria for prostatitis is problematic because a variety of genitourinary conditions could be attributed to prostatitis.
Potential limitations of the present study include misclassification bias in men whose episode of prostatitis did not come to medical attention. Misclassification of control subjects is unlikely because high PSA utilization rates in Olmsted County3 and diagnosed prostate cancer in control subjects would have been identified. Nonetheless, this means that the results represent associations with “diagnosed” prostate cancer. The high intraobserver agreement (see “Methods”) suggests that potential misclassification in ascertainment of exposure (prostatitis) during the abstraction process is minimal. To limit the potential for misclassification bias further, strict prespecified criteria were used for prostatitis. We did not include episodes in the primary analyses that did not meet the criteria such as episodes of chronic symptoms with no evidence of infection or inflammation of the prostate. When questionable episodes were considered in the analyses, the estimates remained essentially the same. Despite these efforts, the diagnosis of chronic prostatitis could be imprecise, limiting the ability to detect associations. Another limitation was the inability to assess the association of sexually transmitted disease with prostate cancer. Most men seek care for sexually transmitted diseases at the Olmsted County Health Clinic, and patient information is not released for research purposes. Finally, the Olmsted County population is primarily a middle class community which, until 1997, was 96% white. The findings can best be generalized to similar settings and to the U.S. white population.
Our findings raise questions about the potential role of infection and inflammation in prostate cancer. The association between acute prostatitis and prostate cancer is intriguing, especially because we found no compelling evidence for the role of chronic prostatitis in prostate cancer. The broad confidence intervals could relate to inadequate sample sizes or to imprecision in the diagnosis of chronic prostatitis. We were limited in our ability to characterize men as having chronic bacterial prostatitis or chronic pelvic pain syndrome (inflammatory type). Biochemical markers such as antibody titers to infectious agents or tissue markers of infection or inflammation could provide better measures and help to clarify the possible role of infection and inflammation in prostate cancer.
We thank Vickey Roeder for abstraction of the data from the medical records and Sondra Buehler for her assistance in the preparation of the manuscript.
1.Jemal A, Thomas A, Murray T, et al. Cancer statistics, 2002. CA Cancer J Clin
2.Roberts RO, Bergstralh EJ, Katusic SK, et al. Decline in prostate cancer mortality from 1980 to 1997, and an update on incidence trends in Olmsted County, Minnesota. J Urol
3.Jacobsen SJ, Katusic SK, Bergstralh EJ, et al. Incidence of prostate cancer diagnosis in the eras before and after serum prostate-specific antigen testing. JAMA
4.Pienta KJ, Esper PS. Risk factors for prostate cancer. Ann Intern Med
5.Whittemore AS, Kolonel LN, Wu AH, et al. Prostate cancer in relation to diet, physical activity, and body size in blacks, whites, and Asians in the United States and Canada. J Natl Cancer Inst
6.Giovannucci E, Rimm EB, Colditz GA, et al. A prospective study of dietary fat and risk of prostate cancer. J Natl Cancer Inst
7.Gann PH, Hennekens CH, Sacks FM, et al. Prospective study of plasma fatty acids and risk of prostate cancer. J Natl Cancer Inst
8.Webber MM, Perez-Ripoll EA, James GT. Inhibitory effects of selenium on the growth of DU-145 human prostate carcinoma cells in vitro. Biochem Biophys Res Commun
9.Combs G Jr, Clark LC. Selenium and cancer. In: Garewel H, ed. Antioxidants and Disease Prevention
. Boca Raton: CRC Press; 1997.
10.Yoshizawa K, Willett WC, Morris SJ, et al. Study of prediagnostic selenium level in toenails and the risk of advanced prostate cancer. J Natl Cancer Inst
11.Heinonen OP, Albanes D, Virtamo J, et al. Prostate cancer and supplementation with alpha-tocopherol and beta-carotene: incidence and mortality in a controlled trial. J Natl Cancer Inst
12.Rohan TE, Howe GR, Burch JD, et al. Dietary factors and risk of prostate cancer: a case-control study in Ontario, Canada. Cancer Causes Control
13.Giovannucci E, Ascherio A, Rimm EB, et al. Intake of carotenoids and retinol in relation to risk of prostate cancer. J Natl Cancer Inst
14.Talamini R, LaVecchia C, Decarli A, et al. Nutrition, social factors and prostatic cancer in a Northern Italian population. Br J Cancer
15.Talamini R, Franceschi S, La Vecchia C, et al. Diet and prostatic cancer: A case-control study in northern Italy. Nutr Cancer
16.Shibata A, Paganini-Hill A, Ross RK, et al. Intake of vegetables, fruits, beta-carotene, vitamin C and vitamin supplements and cancer incidence among the elderly: a prospective study. Br J Cancer
17.Key T. Risk factors for prostate cancer. Cancer Surv
18.Moyret-Lalle C, Marcais C, Jacquemier J, et al. ras, p53 and HPV status in benign and malignant prostate tumors. Int J Cancer
19.Suzuki H, Komiya A, Aida S, et al. Detection of human papillomavirus DNA and p53 gene mutations in human prostate cancer. Prostate
20.Dillner J, Knekt P, Boman J, et al. Sero-epidemiological association between human-papillomavirus infection and risk of prostate cancer. Int J Cancer
21.Griffiths TR, Mellon JK. Human papillomavirus and urological tumours: II. Role in bladder, prostate, renal and testicular cancer. BJU Int
22.Centifanto YM, Kaufman HE, Zam ZS, et al. Herpesvirus particles in prostatic carcinoma cells. J Virol
23.Rosenblatt KA, Wicklund KG, Stanford JL. Sexual factors and the risk of prostate cancer. Am J Epidemiol
24.Mandel JS, Schuman LM. Sexual factors and prostatic cancer: results from a case-control study. J Gerontol
25.Hayes RB, Pottern LM, Strickler H, et al. Sexual behaviour, STDs and risks for prostate cancer. Br J Cancer
26.Ilic M, Vlajinac H, Marinkovic J. Case-control study of risk factors for prostate cancer. Br J Cancer
27.Cuzick J. Human papillomavirus infection of the prostate. Cancer Surv
28.Strickler HD, Schiffman MH, Shah KV, et al. A survey of human papillomavirus 16 antibodies in patients with epithelial cancers. Eur J Cancer Prev
29.Wideroff L, Schottenfeld D, Carey TE, et al. Human papillomavirus DNA in malignant and hyperplastic prostate tissue of black and white males. Prostate
30.Serfling U, Ciancio G, Zhu WY, et al. Human papillomavirus and herpes virus DNA are not detected in benign and malignant prostatic tissue using the polymerase chain reaction. J Urol
31.Effert PJ, Frye RA, Neubauer A, et al. Human papillomavirus types 16 and 18 are not involved in human prostate carcinogenesis: analysis of archival human prostate cancer specimens by differential polymerase chain reaction. J Urol
32.Masood S, Rhatigan RM, Powell S, et al. Human papillomavirus in prostatic cancer: no evidence found by in situ DNA hybridization. South Med J
33.Strickler HD, Burk R, Shah K, et al. A multifaceted study of human papillomavirus and prostate carcinoma. Cancer
34.Ewings P, Bowie C. A case-control study of cancer of the prostate in Somerset and east Devon. Br J Cancer
35.Banerjee AK. Carcinoma of prostate and sexual activity. Urology
36.Krieger JN, Nyberg L Jr, Nickel JC. NIH consensus definition and classification of prostatitis. JAMA
37.Mills PK, Beeson WL, Phillips RL, et al. Cohort study of diet, lifestyle, and prostate cancer in Adventist men. Cancer
38.Nakata S, Imai K, Yamanaka H. Study of risk factors for prostatic cancer [in Japanese]. Acta Urol Jpn
39.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
40.Wynder EL, Mabuchi K, Whitmore WF. Epidemiology of cancer of the prostate. Cancer
41.Mishina T, Watanabe H, Araki H, et al. Epidemiological study of prostatic cancer by matched-pair analysis. Prostate
42.Dennis LK, Lynch CF, Torner JC. Epidemiologic association between prostatitis and prostate cancer. Urology
43.Zhu K, Stanford JL, Daling JR, et al. Vasectomy and prostate cancer: a case-control study in a health maintenance organization. Am J Epidemiol
44.Hiatt RA, Armstrong MA, Klatsky AL, et al. Alcohol consumption, smoking, and other risk factors and prostate cancer in a large health plan cohort in California (United States). Cancer Causes Control
45.Ross RK, Paganini-Hill A, Henderson BE. The etiology of prostate cancer: what does the epidemiology suggest? Prostate
46.Melton LJ III. History of the Rochester Epidemiology Project. Mayo Clin Proc
47.Kurland LT, Molgaard CA. The patient record in epidemiology. Sci Am
48.Krieger JN, Riley DE, Roberts MC, et al. Prokaryotic DNA sequences in patients with chronic idiopathic prostatitis. J Clin Microbiol
49.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
50.Platz EA. Prostatitis and prostate cancer. New Dev Prostate Cancer Treat
51.Baker LH, Mebust WK, Chin TD, et al. The relationship of herpesvirus to carcinoma of the prostate. J Urol
52.Niijima T, Koiso K. Incidence of prostatic cancer in Japan and Asia. Scand J Urol Nephrol
53.Checkoway H, DiFerdinando G, Hulka BS, et al. Medical, life-style, and occupational risk factors for prostate cancer. Prostate