PROSTATE CANCER EARLY DETECTION PRACTICES AMONG MEN WITH A FAMILY HISTORY OF DISEASE CATHRYN H. BOCK, PATRICIA A. PEYSER, STEPHEN B. GRUBER, SUSAN E. BONNELL, KAREN L. TEDESCO, AND KATHLEEN A. COONEY
ABSTRACT Objectives. Genetic studies of familial prostate cancer, which is often asymptomatic until advanced stages, rely on correct designation of affection status. In this pilot study, we set out to determine the proportion of unaffected men whose families are participating in a study of hereditary prostate cancer who have been tested for prostate cancer with serum prostate-specific antigen (PSA) measurement and digital rectal examination (DRE). Methods. Participants were identified from the University of Michigan Prostate Cancer Genetics Project, a family-based study of inherited prostate cancer susceptibility. Of the 141 eligible affected and unaffected sons of men with prostate cancer, 124 (88%) completed a mailed questionnaire regarding serum PSA testing and DRE history. Results. Among unaffected men, 95% reported ever having had a PSA test, and 97% ever having had a DRE, with most initial tests occurring between the ages of 40 and 60 years. No significant difference in the mean age at first PSA test or DRE between the affected and unaffected men was observed. Affected men were significantly more likely than unaffected men to have a first PSA level greater than 2.5 ng/mL (P ⫽ 0.006), but not greater than 4.0 ng/mL (P ⫽ 0.614). Conclusions. Most men with a family history of prostate cancer are undergoing early detection testing. The differences in early detection testing practices do not appear to account for the difference in affection status among the sons of men with prostate cancer. UROLOGY 62: 470–475, 2003. © 2003 Elsevier Inc.
rostate cancer screening recommendations differ among medical organizations. Several advocate routine screening using both prostate-specific antigen (PSA) testing and digital rectal examination (DRE) beginning at age 50 for men with an average risk of prostate cancer and at earlier ages for men at higher risk, including men with a family
This work was supported by National Human Genome Research Institute grant T32 HG00040, S.P.O.R.E. in Prostate Cancer P50 CA69568, National Institutes of Health grants CA79596, CA72818, and CA15083, and the University of Michigan Comprehensive Cancer Center. From the Department of Epidemiology, University of Michigan School of Public Health; Departments of Internal Medicine, Human Genetics, and Urology, University of Michigan Medical School; and Department of Veterans Affairs Medical Center, Ann Arbor, Michigan Reprint requests: Kathleen A. Cooney, M.D., Department of Internal Medicine, University of Michigan Medical School, 7310 CCGC, 1500 East Medical Center Drive, Ann Arbor, MI 481090946 Submitted: January 24, 2003, accepted (with revisions): March 27, 2003
© 2003 ELSEVIER INC. ALL RIGHTS RESERVED
history (FH) of prostate cancer or men who are African American.1–3 Guidelines from several other professional societies, however, do not recommend any routine screening for prostate cancer (Table I).4 – 6 Relatively little is known about prostate cancer early detection testing participation among men at increased risk. Independent statewide surveys of men without a prostate cancer diagnosis estimated that 51% to 64% of men of screening age received PSA tests and up to 77% had undergone a DRE.7–10 Although FH was ascertained in some studies, the participation rates among unaffected men with a positive FH of the disease remain unclear, because the number of subjects with an FH of prostate cancer in each study was generally small. However, in a recent study of Swedish men older than 49 years of age and with three or more relatives with prostate cancer, 68% underwent regular prostate cancer screening,11 and in a Fox Chase Cancer Center study, 88% of men older than 49 years of age with 0090-4295/03/$30.00 doi:10.1016/S0090-4295(03)00382-0
TABLE 1. Current screening guidelines Organization
Annual PSA testing and DRE at age 50 for men at average risk. Men with a first-degree relative diagnosed at a young age or who are African American should begin annual PSA testing and DRE at age 45. Men with multiple affected first-degree relatives diagnosed at an early age should consider starting testing at age 40. American Urological PSA testing and DRE at age 50 for men at Association2 average risk. Men with a strong family history or who are African American should be offered testing earlier. National DRE and offer PSA test at age 50. Every 2 Comprehensive years if prior PSA value ⬍2 ng/mL, annually Cancer Network3 if prior PSA ⱖ2 ng/mL. Annual screening beginning at age 45 for men with a family history or African-American race. U.S. Preventive There is insufficient evidence to recommend Services Task for or against screening. Force4 American College of Routine screening is not recommended; Physicians5 decision to screen should be made on an individual basis. American College of Routine screening is not recommended; Preventive decision to screen should be made on an Medicine6 individual basis. American Cancer Society1
Screening tests should only be offered to men with at least a 10-year life expectancy. Patients should be given information regarding potential risks and benefits of screening tests and treatment.
Decision to test should be made on an individual basis. Potential risks and benefits of testing should be discussed with the patient. Screening tests should only be offered to men with greater than 10-year life expectancy. Risks and benefits of testing should be discussed with the patient. If screening is offered, the patient should be informed of potential risks and harms. Potential risks and benefits of PSA testing and DRE and of treatment options should be discussed with men beginning at age 50. Beginning at age 50, information should be provided regarding the potential benefits and harms of testing, plus the limits of evidence of benefits of screening.
KEY: PSA ⫽ prostate-specific antigen; DRE ⫽ digital rectal examination.
one or more first-degree relatives with prostate cancer had ever had a DRE and 83% had ever had a PSA test.12 Analytic designs used in family-based studies of prostate cancer genetics often include both affected and unaffected individuals. Participating men are usually classified as unaffected only on the basis of self-report, biasing results toward the null because some may have undetected disease. Within prostate cancer genetics studies, it is possible that men with undetected prostate cancer have been misclassified as unaffected because they had not undergone prostate cancer early detection procedures. The Prostate Cancer Screening Study (PCSS) was a pilot study to evaluate prostate cancer early detection practices among men with an FH of prostate cancer who had participated in a prostate cancer genetics study. The primary aim was to determine the proportion of unaffected men who had undergone PSA testing and DRE. Additional aims were to describe the age at the first PSA test and DRE of familial prostate cancer study participants; determine whether affected and unaffected participants differed with respect to age at the first PSA test and DRE and the reasons for the first PSA test and DRE; and determine whether affected and UROLOGY 62 (3), 2003
unaffected men differed with respect to the first PSA value using the standard cutoff values of 2.5 and 4.0 ng/mL. MATERIAL AND METHODS PCSS subjects were selected from families participating in the University of Michigan Prostate Cancer Genetics Project (PCGP). The PCGP was initiated in 1995 to define the molecular basis of hereditary prostate cancer, and PCGP families have been used to examine several prostate cancer candidate loci, including HPC1,13 HPCX,14 and HPC20.15 Fifty-one families with a father with prostate cancer and both one affected and one unaffected living son older than 30 years of age were identified from the PCGP database. All 141 eligible affected and unaffected sons in these families were mailed the PCSS survey, consent form, invitation letter, and a return envelope. All participants provided written informed consent. The Institutional Review Board at the University of Michigan Medical School approved all research protocols. The PCSS survey consisted of self-administered multiple choice and short-answer questions about nonmalignant prostate conditions, PSA testing, and DRE history. Participants were asked to describe the reasons for having their first PSA test and DRE; multiple responses were allowed. Affected men were asked only about PSA tests and DREs that occurred before those that led to their prostate cancer diagnosis. Questions asking for an age when an event occurred allowed men to select an age category (eg, 50 to 59 years), if they could not recall the exact age. Most questions were taken with permission from a questionnaire used in a case-control study of pros471
TABLE II. Prostate cancer early detection practices
Description of study participants Mean age (yr) at survey (⫾ SD) Mean age (yr) at diagnosis (⫾ SD) Race (%) White African American PSA testing* Ever had PSA test† (%) Age (yr) at first PSA test‡ (%) ⬍40 40–49 50–59 ⱖ60 DRE* Ever had DRE test† (%) Age (yr) at first DRE test§ (%) ⬍40 40–49 50–59 ⱖ60
Affected Men (n ⴝ 60)
Unaffected Men (n ⴝ 64)
58.9 (⫾ 9.21) 54.1 (⫾ 8.08)
55.6 (⫾ 10.34) NA
55 (91.7) 5 (8.3)
61 (95.3) 3 (4.7)
2 21 17 3
10 19 18 13
(4.7) (48.8) (39.5) (7.0)
(16.7) (31.7) (30.0) (21.7)
16 23 11 2
17 22 13 8
(30.8) (44.2) (21.2) (3.8)
(28.3) (36.7) (21.7) (13.3)
Abbreviations as in Table I. * Among affected men, PSA or DRE before the one that led to prostate cancer diagnosis; remaining affected men not included because first PSA test and/or DRE led to diagnosis of prostate cancer. † One affected man did not report. ‡ One unaffected man did not report. § Two affected and two unaffected men did not report.
tate cancer in African-American men.16 The questionnaire took approximately 20 minutes to complete; subjects were not compensated for participation. Medical records obtained by the PCGP provided information regarding PSA levels and tumor stage for the affected men, and tumors were classified using the Surveillance, Epidemiology, and End Results criteria.17 Data were analyzed using the Statistical Package for Social Sciences, version 10.0.5 (SPSS, Chicago, Ill) and Statistical Analysis System, version 8.01 (SAS Institute, Cary, NC) statistical software. Generalized estimating equations18 implemented under the exchangeable covariance structure, adjusting for intrafamilial correlation, were used to compare the affected and unaffected men. All tests were two-sided, and P values of 0.05 or less were considered statistically significant.
RESULTS Of 141 eligible men, 124 (88%) from all 51 families returned a completed survey. The response rate was 94% (60 of 64) for affected men and 83% (64 of 77) for unaffected men. Participants did not differ significantly from nonparticipants with respect to age at the time of the study, age at diagnosis among affected men, or race (data not shown). Eight men (6%) from four families were African American; the remaining 116 men (94%) from 47 families were white. The average age was 59 years (range 38 to 84, median 57) among those with a prostate cancer diagnosis and 56 years (range 31 to 78, median 55) among those without (Table II). Significantly more affected than unaffected men re472
ported a diagnosis of prostatitis (P ⫽ 0.024). No difference was found, however, in benign prostatic hyperplasia diagnosis (P ⫽ 0.822). At the time of prostate cancer diagnosis, one affected man (1.7%) had evidence of distant metastatic disease; the remainder had localized (73.3%) or regional (25.0%) prostate cancer. Among the unaffected men, 95.3% had ever had a PSA test, 96.9% had ever had a DRE (Table II), 93.8% had had both, 1.6% had had a PSA test only, and 3.1% had had only a DRE (data not shown). Of the 60 affected men, 71.7% had had a PSA test before the one that led to a prostate cancer diagnosis (Table II), 91.5% had had a DRE prior to diagnosis, and 70.0% had had both PSA testing and a DRE (data not shown). Among the 16 affected men whose first PSA test led to the diagnosis of prostate cancer, diagnostic PSA values were available for 13 men; in 12 of these cases, the PSA value was greater than 4.0 ng/mL. Collectively, most participants received their first PSA test between the ages of 40 and 60 years, and approximately one half received their first PSA test before age 50 (Table II). All but one of the initial PSA tests occurred before the age of 70 years. No statistically significant difference in age at the first PSA test was noted between the affected and unaffected men. The most common reasons cited by both affected and unaffected men for undergoUROLOGY 62 (3), 2003
TABLE III. Comparison of first PSA values among affected and unaffected men First PSA Value (ng/mL) ⬎2.5* (%) ⬎4.0* (%)
Affected Men (n ⴝ 27)
Unaffected Men (n ⴝ 28)
14 (51.9) 3 (11.1)
4 (14.3) 2 (7.1)
KEY: PSA ⫽ prostate-specific antigen. Numbers in parentheses are percentages. * Among affected men, PSA before one that led to prostate cancer diagnosis.
ing an initial PSA test were that the patient (39.4%) and/or his physician (35.6%) requested it. An FH of prostate cancer (31.7%) was also frequently cited. When comparing the two groups, unaffected men were more likely to report an FH of prostate cancer as a reason for their first PSA test (P ⫽ 0.022). Most initial DREs took place before the age of 50, and all occurred before age 70 (Table II). No differences were noted between the two groups with regard to the proportion ever having had a DRE or the age at the first DRE. Within both groups, the most common reason given for having an initial DRE was physician request (37.1%). Having a DRE as part of a checkup (26.7%) or because the patient requested it (23.3%) were also often reported as reasons for an initial DRE. No statistically significant differences were found between the affected and unaffected men in the reasons given for a first DRE. Initial PSA values were reported by 27 of the affected (62.8%) and 28 of the unaffected (45.9%) men who indicated having ever had a PSA test (Table III). No statistically significant difference was found between the two groups in the proportion of men with a first PSA value greater than 4.0 ng/mL. Affected men, however, were significantly more likely to have a first PSA value greater than 2.5 ng/mL (P ⫽ 0.006) compared with unaffected men. COMMENT The proportion of men undergoing prostate cancer early detection testing was greater than 95%, suggesting that the potential for misclassification of affected men as unaffected is low in this study group. If the nonresponders never had a PSA test, a conservative estimated proportion of unaffected men who underwent PSA testing would be 79.2% (61 of 77). If these findings can be generalized to all men participating in familial prostate cancer studies, it is unlikely that misclassification of affected men as unaffected was introducing a strong bias. The proportions of unaffected participants who reported a history of PSA testing or DRE were approximately 20% greater than those reported UROLOGY 62 (3), 2003
among men in the general population7–10 and approximately 10% greater than those reported among men with an FH of prostate cancer.11,12 Our participants were likely more motivated to undergo prostate cancer early detection testing than men surveyed from the general population because the PCSS subjects were recruited from a group of men with an FH of prostate cancer who had also participated in a genetics study. The observed proportions may reasonably be generalized to participants in other familial prostate cancer studies, but are probably an overestimate of the proportions in the population of all men with an FH of prostate cancer. At least one half of respondents (affected and unaffected) met American Cancer Society guidelines for prostate cancer screening in men with an FH of prostate cancer, which recommend initial PSA testing and DRE beginning at age 45.1 Some men might have been first tested at later ages, because the current guidelines did not exist when the men were younger. Also, some men did not have an FH of disease until a later age. The earlier age at the first DRE compared with the first PSA test may be both because of the historically later introduction of the PSA test and because the DRE is a screening procedure for other disorders. Similar testing patterns between affected and unaffected men are suggested by the lack of statistically significant differences in age at the first PSA test and age at the first DRE. Few men reported that their initial PSA test or DRE occurred for clinical reasons, such as symptoms or an abnormal test result. Thus, most participants probably underwent initial early detection tests for screening purposes. Unaffected men were more likely to report an FH of prostate cancer as their reason compared with affected men, and this difference may be attributed to selection bias introduced by the study design. However, this bias should not weaken our conclusion that affection status differences were not due to variation in prostate cancer early detection practices. Age differences between the affected and unaffected men at the time of the study were expected, because prostate cancer risk increases with age, 473
and therefore the men with prostate cancer were more likely to have an older mean age than their unaffected siblings. Including age at participation as a covariate did not change any inferences (data not shown). Disagreement exists regarding the appropriate serum PSA cutpoint values for prostate cancer screening because of the uncertainty regarding the potential benefits of PSA testing combined with the known risks arising from false-positive test results.19 Compelling arguments exist for the use of cutoff values of either 4.0 ng/mL20 or 2.5 ng/mL.21 In this pilot study, the observation of statistically significant differences between the proportion of affected and unaffected men having a self-reported first PSA value greater than 2.5 ng/mL, but not greater than 4.0 ng/mL, suggests that additional investigation of the most appropriate cutpoint is warranted among men with an FH of prostate cancer. This suggestion is supported by the results from a study of men with initial PSA values between 2.5 and 4.0 ng/mL22 in which a greater proportion of men with positive biopsy results had an FH of prostate cancer (34%) compared with men with a negative report (22%), although this difference was not statistically significant.22 Future studies of PSA cutpoints for men with an FH of prostate cancer should use medical records to determine the PSA values to avoid biases introduced by selfreport in PSA level, to determine whether a biopsy was performed, and to compare the free PSA values of affected and unaffected men. A potential problem in the PCSS is the use of self-reported rather than medical record information. Several studies have examined the agreement of self-report of DRE with the information in medical records and generally observed low to moderate levels of agreement between the two sources (kappa values ranged from ⫺0.05 to 0.54).23–25 One study also compared self-reported PSA values with medical records within the past 2 years and found a moderate level of agreement (kappa 0.38).25 However, documentation of PSA and DRE tests may be difficult because of the variety of clinical settings in which these tests may be performed and missing data. The reported kappa values probably represent an overly conservative estimation of agreement between self-reported and actual PSA testing and DRE history. Our study included prevalent rather than incident cases, and affected men at the time of the survey were almost 5 years older than their age at diagnosis. This could have introduced a survival bias because men with more severe disease have a shorter expected survival time18 and would be less likely to be available for study inclusion. However, the distribution of tumor stages in our study is 474
consistent with that expected if only cases with new diagnoses were sampled.26 CONCLUSIONS The most appropriate serum PSA cutpoint value among men with an FH of prostate cancer warrants further investigation. Most men with an FH of prostate cancer are undergoing early detection testing, and differences in early detection testing practices do not appear to account for the difference in affection status among sons of men with prostate cancer. ACKNOWLEDGMENT. To Drs. Gary J. Faerber and David P. Wood for critical review of this manuscript. REFERENCES 1. Smith RA, von Eschenbach AC, Wender R, et al: American Cancer Society guidelines for the early detection of cancer— update of early detection guidelines for prostate, colorectal, and endometrial cancers and update 2001: testing for early lung cancer detection. CA Cancer J Clin 51: 38 –75, 2001. 2. Carroll P, Coley C, McLeod D, et al: Prostate-specific antigen best practice policy—part I: early detection and diagnosis of prostate cancer. Urology 57: 217–224, 2001. 3. Potter SR, and Partin A: NCCN practice guidelines for early detection of prostate cancer. Oncology 13: 99 –115, 1999. 4. U.S. Preventive Services Task Force: Screening for prostate cancer: recommendation and rationale. Ann Intern Med 137: 915–916, 2002. 5. Coley CM, Barry MJ, Fleming C, et al: Early detection of prostate cancer, part II: estimating the risks, benefits, and costs. Ann Intern Med 126: 468 –479, 1997. 6. Ferrini R, and Woolf SH: Screening for prostate cancer in American men: American College of Preventive Medicine Practice Policy Statement. Am J Prev Med 15: 81–84, 1988. 7. McDavid K, Melnik TA, and Derderian H: Prostate cancer screening trends of New York State men at least 50 years of age, 1994 to 1997. Prev Med 31: 195–202, 2000. 8. Michigan Public Health Institute: Screening for Prostate and Colorectal Cancer: Findings from Surveys of Adults and Primary Care Physicians in Michigan—Volume I: Prostate Cancer. Ann Arbor, Michigan Public Health Institute, 1997. 9. Close DR, Kristal AR, Li S, et al: Associations of demographic and health-related characteristics with prostate cancer screening in Washington state. Cancer Epidemiol Biomarkers Prev 7: 627–630, 1998. 10. Lemon S, Zapka J, Puleo E, et al: Colorectal cancer screening participation: comparisons with mammography and prostate-specific antigen screening. Am J Pub Health 91: 1264 –1272, 2001. 11. Bratt O, Damber J-E, Emanuelsson M, et al: Risk perception, screening practice and interest in genetic testing among unaffected men in families with hereditary prostate cancer. Eur J Cancer 36: 235–241, 2000. 12. Miller SM, Diefenbach MA, Kruus LK, et al: Psychological and screening profiles of first-degree relatives of prostate cancer patients. J Behav Med 24: 247–258, 2001. 13. Cooney KA, McCarthy JD, Lange E, et al: Prostate cancer susceptibility locus on chromosome 1q: a confirmatory study. J Natl Cancer Inst 89: 955–959, 1997. 14. Lange EM, Chen H, Brierley K, et al: Linkage analysis of 153 prostate cancer families over a 30-cM region containing UROLOGY 62 (3), 2003
the putative susceptibility locus HPCX. Clin Cancer Res 5: 4013–4020, 1999. 15. Bock CH, Cunningham JM, McDonnell SK, et al: Analysis of the prostate cancer-susceptibility locus HPC20 in 172 families affected by prostate cancer. Am J Hum Genet 68: 795–801, 2001. 16. Cooney KA, Strawderman MS, Wojno KJ, et al: Agespecific distribution of serum prostate-specific antigen in a community-based study of African-American men. Urology 57: 91–96, 2001. 17. Stanford JL, Stephenson RA, Coyle LM, et al: Prostate Cancer Trends 1973–1995, NIH Publication No. 99-4543. Bethesda, National Institutes of Health, 1999. 18. Liang KY, and Zeger SL: Longitudinal data analysis using generalized linear models. Biometrika 73: 13–22, 1986. 19. Harris R, Lohr KN, Beck R, et al: Screening for prostate cancer: systematic evidence review No. 16. Rockville, Maryland, Agency for Healthcare Research and Quality, www.ahrq. gov/clinic/serfiles.htm, 2002. Accessed February 26, 2003. 20. Carter HB: A PSA threshold of 4.0 ng/mL for early detection of prostate cancer: the only rational approach for men 50 years old and older. Urology 55: 796 –799, 2000.
UROLOGY 62 (3), 2003
21. Catalona WJ, Ramos CG, Carvalhal GF, et al: Lowering PSA cutoffs to enhance detection of curable prostate cancer. Urology 55: 791–795, 2000. 22. Babaian RJ, Johnston DA, Naccarato W, et al: The incidence of prostate cancer in a screening population with a serum prostate specific antigen between 2.5 and 4.0 ng/ml: relation to biopsy strategy. J Urol 165: 757–760, 2001. 23. Gordon NP, Hiatt RA, and Lampert DI: Concordance of self-reported data and medical record audit for six cancer screening procedures. J Natl Cancer Inst 85: 566 –570, 1993. 24. Jordan TR, Price JH, King KA, et al: The validity of male patients’ self-reports regarding prostate cancer screening. Prev Med 28: 297–303, 1999. 25. Zhu K, McKnight B, Stergachis A, et al: Comparison of self-report data and medical records data: results from a casecontrol study on prostate cancer. Int J Epidemiol 28: 409 – 417, 1999. 26. Ries LAG, Eisner MP, Kosary CL, et al (Eds): SEER cancer statistics review, 1973–1999. Bethesda, National Cancer Institute, http://seer.cancer.gov/csr/1973_1999/, 2002. Accessed April 15, 2002.