Portal de Periódicos da CAPES

Prostate-specific antigen testing for early diagnosis of prostate cancer: formulation of guidelines

1999; Elsevier BV; Volume: 54; Issue: 5 Linguagem: Inglês

10.1016/s0090-4295(99)00271-x

1527-9995

H.Ballentine Carter, Jay D. Pearson,

Colorectal Cancer Screening and Detection

Population trends before and after the onset of the prostate-specific antigen (PSA) era suggest that PSA testing has some of the characteristics associated with a screening program that reduces disease-specific mortality. Although PSA testing is widely used to diagnose prostate cancer in the United States at earlier stages (case finding), efforts to define PSA testing guidelines have just begun. Guidelines determine the burden (financial and personal) of testing for a population and should be addressed as information from serial PSA testing of various populations becomes available. Criteria regarding the age for PSA testing to begin, the interval between PSA testing, and the age at which PSA testing can be discontinued or relaxed should be defined in the context of the goal of detection of curable prostate cancer in those who will benefit from treatment. In this article, no attempt is made to review the unresolved controversy surrounding the routine use of digital rectal examination and PSA testing in asymptomatic men or the performance of prostate cancer screening tests, both of which have been addressed by others.1Barry M.J Fleming C Coley C.M et al.Should Medicare provide reimbursement for prostate-specific antigen testing for early detection of prostate cancer? Part I framing the debate.Urology. 1995; 46: 2-13Abstract Full Text PDF Scopus (31) Google Scholar, 2Kramer B.S Brown M.L Prorok P.C et al.Prostate cancer screening what we know and what we need to know.Ann Intern Med. 1993; 119: 914-923Crossref Scopus (211) Google Scholar, 3Coley C.M Barry M.J Fleming C et al.Early detection of prostate cancer. Part I: prior probability and effectiveness of tests.Ann Intern Med. 1997; 126: 394-406Crossref Scopus (161) Google Scholar, 4Middleton R.G Prostate cancer are we screening and treating too much?.Ann Intern Med. 1997; 126: 465-467Crossref Scopus (12) Google Scholar Instead, this article attempts to review some of the initial efforts to define appropriate guidelines for PSA testing with the assumption that PSA testing will continue to be widespread in the population. Widespread PSA testing for the detection of prostate cancer began in the late 1980s and resulted in a dramatic increase in the incidence of prostate cancer that peaked in 1992.5Stephenson R.A Smart C.R Mineau G.P et al.The fall in incidence of prostate carcinoma on the downside of a prostate specific antigen induced peak in incidence-data from the Utah Cancer Registry.Cancer. 1996; 77: 1342-1348Crossref Scopus (93) Google Scholar This increase in disease incidence—and subsequent increase in treatment for prostate cancer—has fueled the controversy regarding the use of PSA testing. The lack of randomized trials demonstrating reductions in prostate cancer mortality with PSA testing constitute the strongest argument for restraint. However, there is growing evidence that PSA testing may save lives. A successful screening program for cancer (as exemplified by cervical cancer screening) results in a decrease in cancer-specific mortality of the disease in question and has certain characteristics. First, a test must be available that results in an increase in detection lead time. Second, after introduction of the test into the population, the incidence of the disease should initially increase (if the disease is being detected earlier) as prevalent cases that would have gone undetected are removed from the population (cull effect). Third, the incidence of advanced disease should be reduced at some point after the introduction of the test. Fourth, a reduction in mortality should occur, and this reduction should be related to the intensity of screening, with greater decreases in mortality among the population more intensively screened. It has been demonstrated in PSA-based screening studies that, in comparison to a referral population, men diagnosed with prostate cancer by PSA testing have a higher probability of confined cancer.6Catalona W.J Smith D.S Ratliff T.L et al.Detection of organ-confined prostate cancer is increased through prostate-specific antigen based screening.JAMA. 1993; 270: 948-954Crossref Scopus (868) Google Scholar The increase in detection of confined prostate cancer with PSA testing is a result of the increased lead time for detection compared with detection by other methods (eg, digital rectal examination). The increase in detection lead time that results with PSA testing has been estimated to be approximately 5 years—a value between the mean lead times associated with mammography (1.7 years) and cervical cytology (10 to 20 years).7Gann P.H Hennekens C.H Stampfer M.J A prospective evaluation of plasma prostate-specific antigen for detection of prostatic cancer.JAMA. 1995; 273: 289-294Crossref Scopus (627) Google Scholar, 8Carter H.B Pearson J.D PSA and the natural course of prostate cancer.in: Schroder F.H Recent Advances in Prostate Cancer and BPH. Parthenon Publishing Group, New York1997: 187-193Google Scholar One would anticipate an increase in disease incidence after the introduction of a test that increases lead time, because prevalent cases are removed from the population (cull phenomenon). The increase in prostate cancer incidence after the introduction of widespread PSA testing5Stephenson R.A Smart C.R Mineau G.P et al.The fall in incidence of prostate carcinoma on the downside of a prostate specific antigen induced peak in incidence-data from the Utah Cancer Registry.Cancer. 1996; 77: 1342-1348Crossref Scopus (93) Google Scholar, 9Jacobsen S.J Katusic S.K Bergstralh E.J et al.Incidence of prostate cancer diagnosis in the eras before and after serum prostate-specific antigen testing.JAMA. 1995; 274: 1445-1449Crossref Scopus (199) Google Scholar is analogous to the increase in cervical cancer incidence that occurred after the introduction of successful screening programs for cervical cancer.10Sigurdsson K Effect of organized screening on the risk of cervical cancer. Evaluation of screening activity in Iceland, 1964–1991.Int J Cancer. 1993; 54: 563-570Crossref Scopus (106) Google Scholar Furthermore, decreases in the rate of advanced prostate cancer have been documented after the onset of widespread PSA testing11Stephenson R.A Population-based prostate cancer trends in the PSA era data from the Surveillance, Epidemiology, and End Results (SEER) program.Monogr Urol. 1998; 19: 3-19Google Scholar similar to the decreases in invasive cervical cancer that occurred after the onset of widespread cervical cancer screening.12Cramer D.W The role of cervical cytology in the declining morbidity and mortality of cervical cancer.Cancer. 1974; 34: 2018-2027Crossref Scopus (141) Google Scholar An inverse relationship between disease-specific mortality and the intensity of screening was noted with cervical cancer screening programs in different regions of the United States12Cramer D.W The role of cervical cytology in the declining morbidity and mortality of cervical cancer.Cancer. 1974; 34: 2018-2027Crossref Scopus (141) Google Scholar and throughout Europe,10Sigurdsson K Effect of organized screening on the risk of cervical cancer. Evaluation of screening activity in Iceland, 1964–1991.Int J Cancer. 1993; 54: 563-570Crossref Scopus (106) Google Scholar and these data, in part, serve to support the rationale for the use of regular Papanicolaou smears. A similar relationship between the intensity of PSA testing and prostate cancer mortality in different regions of the United States has yet to be demonstrated. However, prostate cancer mortality in the United States is decreasing,13Wingo P.A Ries L.A.G Rosenberg H.M et al.Cancer incidence and mortality, 1973–1995. A report card for the U.S.Cancer. 1998; 82: 1197-1207Crossref Scopus (374) Google Scholar with the greatest decreases reported among younger white men (a group more intensively involved in early detection and treatment efforts before the PSA era).14Harlin L Brawley O Pommerenke F et al.Geographic, age, and racial variation in the treatment of local/regional carcinoma of the prostate.J Clin Oncol. 1995; 13: 93-100Google Scholar Roberts et al.15Roberts R.O Bergstralh E.J Katusic S.K et al.Decline in prostate cancer mortality from 1980 to 1997, and an update on incidence trends in Olmsted County, Minnesota.J Urol. 1999; 161: 529-533Abstract Full Text Full Text PDF Scopus (108) Google Scholar reported a 22% decline in prostate cancer mortality in Olmsted County, Minnesota (a region with intensive screening) when comparing the eras 1980 to 1984 and 1993 to 1997. Decreases in mortality reported within the first decade after the onset of widespread PSA testing are unlikely to be due to PSA testing given the long natural history of prostate cancer. However, a decline in prostate cancer mortality could be due to the increased interest in the early detection of prostate cancer with digital rectal examination that began before the PSA era, as manifested by the earlier stage at diagnosis (stage migration)16Gilliland F Becker T.M Smith A et al.Trends in prostate cancer incidence and mortality in New Mexico are consistent with an increase in effective screening.Cancer Epidemiol Biomarkers Prev. 1994; 3: 105-111Google Scholar and increasing surgery rates for localized prostate cancer in the decade before the onset of widespread PSA testing.17Mettlin C Murphy G.P Menck H Trends in treatment of localized prostate cancer by radical prostatectomy observations from the Commission on Cancer National Cancer Database, 1985–1990.Urology. 1994; 43: 488-492Abstract Full Text PDF Scopus (49) Google Scholar Jacobsen et al.18Jacobsen S.J Bergstralh E.J Katusic S.K et al.Screening digital rectal examination and prostate cancer mortality a population-based case-control study.Urology. 1998; 52: 173-179Abstract Full Text Full Text PDF Scopus (63) Google Scholar recently reported a strong inverse association between the digital rectal examination and prostate cancer mortality in the pre-PSA era, suggesting that future trends may demonstrate even further decreases in prostate cancer mortality with widespread PSA testing. Demonstration of an inverse relationship between prostate cancer mortality and the intensity of PSA testing in the United States and Europe may occur before data from randomized screening trials are available. If so, the same information that is used to support the value of cervical cancer screening will be available to support prostate cancer screening (see above). The focus of screening discussions will then shift from the appropriateness of PSA testing to appropriate guidelines that result in the most cost-effective methods for early detection of prostate cancer. Guidelines determine the burden of screening in the population in terms of unnecessary tests, false-positive tests, and the downstream effects of false-positive testing and are thus an important aspect of a successful screening program. These guidelines should address the age at which PSA testing should start, the appropriate interval for testing, and a recommended age at which PSA testing should cease or become less intense. Although general guidelines have been suggested by the American Cancer Society and the American College of Physicians,19von Eschenbach A Ho R Murphy G.P et al.American Cancer Society guidelines for the early detection of prostate cancer.Cancer. 1997; 80: 1805-1807Crossref Google Scholar, 20Coley C.M Barry M.J Fleming C et al.Clinical guideline part III screening for prostate cancer.Ann Intern Med. 1997; 126: 480-484Crossref Google Scholar efforts to define specific parameters for PSA testing have just begun. A recommendation has been made to offer PSA testing to men beginning at age 50, and for African Americans and those with a familial predisposition to begin at a younger age (eg, 45 years).19von Eschenbach A Ho R Murphy G.P et al.American Cancer Society guidelines for the early detection of prostate cancer.Cancer. 1997; 80: 1805-1807Crossref Google Scholar, 20Coley C.M Barry M.J Fleming C et al.Clinical guideline part III screening for prostate cancer.Ann Intern Med. 1997; 126: 480-484Crossref Google Scholar This recommendation is based on the increasing incidence of disease after age 50, and the possibility of an earlier age of onset among African Americans and those with a family history. Although this recommendation may be valid, modification may be necessary as more information is acquired. For example, it may be more cost effective to routinely test all men before age 50 (eg, age 40 or 45) and select those men who after an initial screening require more frequent testing on the basis of their baseline PSA level. In a recent study, younger age was clearly demonstrated to be a strong predictor of curable prostate cancer.21Carter H.B Epstein J.I Partin A.W The influence of age and prostate specific antigen on the chance of curable prostate cancer among men with non-palpable disease.Urology. 1999; 53: 126-130Abstract Full Text Full Text PDF Scopus (106) Google Scholar Four hundred ninety-two consecutive men with Stage T1c (nonpalpable) prostate cancer who underwent radical prostatectomy were evaluated for the presence of curable prostate cancer (organ confined or low-grade capsular penetration with negative surgical margins). Only a minimal difference (2% to 4%) was found in the probability of curable cancer at PSA levels between 2.5 and 4.0 ng/mL compared with PSA levels between 4.1 and 6.0 ng/mL for men in three age decades (40 to 50, 51 to 60, and 61 to 73 years). However, there was a more substantial difference in the probability of curable cancer when comparing the three age decades for a given PSA level, with younger men more likely to have curable disease (Table I). These data suggest that a more substantial increase in the detection of curable disease would occur by beginning PSA testing at a younger age than by lowering the PSA cutoff value (the level that would initiate a prostate biopsy) to less than 4.0 ng/mL. Although these data are preliminary, they are consistent with the slow natural history of most prostate cancers, which are thought to take decades to progress to an incurable stage.TABLE IPercentage of probability of curable prostate cancer by age and pretreatment PSA level legendKey: PSA = prostate-specific antigen., legendProbability is given as a percentage (with 95% confidence interval within parentheses)., legendP 10.073 (65–80)57 (52–62)49 (45–53)legend Key: PSA = prostate-specific antigen.legend Probability is given as a percentage (with 95% confidence interval within parentheses).legend P <0.05 for all comparisons. Open table in a new tab Before recommendations can be made to change the age at which PSA testing is initiated, it is necessary to understand the potential cost and benefit of initiating screening at different ages. In the absence of randomized trials comparing different screening strategies, computer simulations of the natural history of disease progression can be used to address rational screening strategies.22Shun-Zhang Y Miller A.B Sherman G.J Optimising the age, number of tests, and test interval for cervical screening in Canada.J Epidemiol Comm Health. 1982; 36: 1-10Crossref Scopus (34) Google Scholar Ross et al.23Ross K.S Pearson J.D Guess H.A et al.Rational PSA screening strategies for prostate cancer computer simulations of age ranges, PSA cutoffs, and testing intervals (abstract).J Urol. 1999; 161: 210Crossref Google Scholar compared different screening strategies using computer simulations of the natural history of prostate cancer progression in a population of men followed up from age 40 to age 80. Different prostate cancer states (in terms of volume and curability of disease), and transitions between states, were assigned on the basis of the prevalence of disease as determined by autopsy and reported prostate cancer mortality rates. PSA distributions for the population with and without prostate cancer and biopsy detection rates were taken from published reports. The investigators found that when compared with annual testing beginning at age 50, a 2-year PSA testing interval beginning at age 40 reduced prostate cancer deaths, the number of PSA tests per curable cancer, and the number of biopsies per curable cancer. In contrast, lowering the PSA cutoff to 3.0 ng/mL, or 2.5 ng/mL, did not prevent more prostate cancer deaths but doubled the number of biopsies per curable cancer. Thus, although yearly testing beginning at age 50 is a general recommendation, it may not be the most cost-effective approach for early detection of prostate cancer. Further study is needed to identify the age at which PSA testing should begin, and this age may be different depending on a patient’s risk factors for the development of prostate cancer. The major focus of clinical research in the area of PSA testing has been on identifying methods (PSA density, age-adjusted PSA values, PSA velocity, and free PSA) for reducing false-positive tests (increasing specificity) while maintaining cancer detection (sensitivity). These efforts seek to balance the tradeoffs between missing an important cancer and the unnecessary evaluations (prostate ultrasound and biopsy) that occur in men with false-positive tests. Relatively little attention has been paid to the predictive value and the natural history of PSA values that are lower than “normal,” information that is important for the development of PSA testing guidelines. It has been shown that the serum PSA level is a strong predictor of a future diagnosis of prostate cancer—even when levels are below “normal.”7Gann P.H Hennekens C.H Stampfer M.J A prospective evaluation of plasma prostate-specific antigen for detection of prostatic cancer.JAMA. 1995; 273: 289-294Crossref Scopus (627) Google Scholar Using a single serum sample collected 10 years before disease ascertainment, Gann et al.7Gann P.H Hennekens C.H Stampfer M.J A prospective evaluation of plasma prostate-specific antigen for detection of prostatic cancer.JAMA. 1995; 273: 289-294Crossref Scopus (627) Google Scholar demonstrated a twofold risk of a prostate cancer diagnosis in men with PSA levels of 1.01 to 1.50 ng/mL and a fivefold risk in men with PSA levels of 2.01 to 3.00 ng/mL compared with men with PSA levels of 1.0 ng/mL or less. The investigators pointed out that dichotomization of PSA values into normal and abnormal could obscure the information inherent in values below the standard cutoff. A serum bank from a prospective aging study (Gerontology Research Center of the National Institute on Aging, the Baltimore Longitudinal Study of Aging)24Shock NW, Greulich RC, Andres R, et al: Normal human aging: the Baltimore Longitudinal Study of Aging. NIH publication No. 84-2450. Washington, DC, United States Government Printing Office, November 1984.Google Scholar has allowed evaluation of repeated PSA measures during two decades among men with and without prostate cancer. Serum PSA levels were measured on a total of 963 men either at the time of routine subject visits (since 1991) or using a frozen serum bank for retrospective samples (before 1991). All PSA measurements were performed using a monoclonal immunoradiometric assay (Tandem-R, Hybritech, San Diego, Calif). This unique resource has provided information regarding the natural history of PSA (Fig. 1). The PSA history of men with and without prostate cancer is one of a slowly progressive rise, even among men who will later be diagnosed with prostate cancer (Fig. 1). These data suggest that a low PSA level is predictive of future low levels and that the intensity of PSA testing can be linked to the baseline PSA level, with less intensive screening in the setting of low baseline levels. This concept is analogous to relaxation of screening intensity among women with negative cervical cancer screens, whereby a negative screen, especially multiple negative screens, is predictive of future negative tests.25IARC Working Group on Evaluation of Cervical Cancer Screening ProgrammesScreening for squamous cervical cancer duration of low risk after negative results of cervical cytology and its implication for screening policies.BMJ. 1986; 293: 659-664Crossref Scopus (448) Google Scholar Annual PSA testing has remained the standard practice since the advent of widespread use of the PSA test for early detection of prostate cancer. However, annual testing may not necessarily lead to improved outcomes with respect to the early detection of prostate cancer at a curable stage when compared with less intensive testing. For example, annual cervical cancer screens do not lead to a significantly greater reduction in the rate of invasive cervical cancer compared with less intensive screening. It has been estimated that yearly screening from age 20 to 64 reduces the rate of invasive cervical cancer by 93%; screening at 3-year intervals from age 20 to 64 reduces the rate of invasive cervical cancer by 91%.26Eddy D.M Screening for cervical cancer.Ann Intern Med. 1990; 113: 214-226Crossref Scopus (396) Google Scholar The outcome of interest (the percentage of reduction in the rate of invasive cervical cancer) is little changed with markedly different screening intervals. This paradigm may be applicable to PSA testing. Carter et al.27Carter H.B Epstein J.I Chan D.W et al.Recommended prostate-specific antigen testing intervals for detection of curable prostate cancer.JAMA. 1997; 277: 1456-1460Crossref Google Scholar evaluated PSA testing intervals that would maintain the detection of curable prostate cancer using the data from the Baltimore Longitudinal Study of Aging and a case series of men treated surgically for nonpalpable prostate cancer. First, the PSA cutoff that would maintain the detection of curable prostate cancer and minimize the detection of small cancers was determined by evaluation of 389 consecutive men with Stage T1c prostate cancer who underwent radical prostatectomy. Significant tumors were defined as all tumors, except small cancers (organ confined, no Gleason pattern 4 or 5, and volume less than 0.5 cm3). Noncurable cancers were advanced tumors (capsular penetration of high grade [Gleason score 7 or greater], established [nonfocal] capsular penetration of any grade with positive margins, or involvement of seminal vesicles or lymph nodes) with PSA relapse-free rates of less than 50% a decade after surgery. Curable cancers were those without the pathologic characteristics of advanced tumors associated with PSA relapse-free rates a decade after surgery of more than 75%.27Carter H.B Epstein J.I Chan D.W et al.Recommended prostate-specific antigen testing intervals for detection of curable prostate cancer.JAMA. 1997; 277: 1456-1460Crossref Google Scholar When the pretreatment PSA was 4.0 ng/mL or less, nonpalpable prostate cancers were highly likely (94%) to be curable, and most (69%) were small cancers that were potentially unimportant. When the pretreatment PSA level was greater than 4.0 ng/mL but 5.0 ng/mL or less, cancers were highly likely to be curable (89%), and a minority (33%) were small cancers. When the pretreatment PSA level was greater than 5.0 ng/mL, 30% of the cancers were noncurable. Figure 2 summarizes these data and emphasizes the tradeoffs inherent with PSA testing (ie, a higher probability of insignificant cancer at low PSA levels and a lower probability of curable cancer at higher PSA levels). These data suggest that a PSA cutoff of 4.0 ng/mL is valid for maintaining the detection of curable disease. Second, the probability of a PSA conversion to a level of 4.0 ng/mL and to greater than 5.0 ng/mL was determined using the Baltimore Longitudinal Study of Aging data base of serial PSA values in 40 men who eventually developed prostate cancer and 272 men without cancer (Fig. 3). The PSA conversion (for cancer cases) to a level at which cure was less likely (greater than 5.0 ng/mL) was rare (0%) after 2 and 4 years when the initial PSA was less than 2.0 ng/mL. The PSA conversion to a range at which cancers were likely to be curable and less likely to be small (4.1 to 5.0 ng/mL) was rare after 2 years (0% to 4%) when the baseline PSA level was less than 2.0 ng/mL but common when the baseline PSA level was between 2.1 and 3.0 ng/mL (27%) or 3.1 and 4.0 ng/mL (36%). Taken together, these data suggest that a 2-year PSA testing interval is not likely to miss a curable prostate cancer among men with no palpable suspicion of cancer on digital rectal examination and a PSA level less than 2.0 ng/mL. For men at higher risk of the development of prostate cancer (African American or family history of the disease), more intensive yearly PSA testing may be appropriate. Two studies support the above recommendation for a 2-year PSA testing interval for most men with PSA levels less than 2.0 ng/mL. First, data from the Washington University screening experience28Smith D.S Catalona W.J Herschman J.D Longitudinal screening for prostate cancer with prostate specific antigen.JAMA. 1996; 276: 1309-1315Crossref Google Scholar demonstrate that after 4 years of testing 8662 men at 6-month intervals, the rate of PSA conversion to more than 4.0 ng/mL was 4% when the baseline PSA was less than 2.5 ng/mL and 48% when the baseline PSA level was 2.6 to 4.0 ng/mL. Second, data from Prostate Cancer Awareness Week29Leewansangtong S Crawford E.D Gordon S.G et al.Longitudinal follow up from Prostate Cancer Awareness Week (PCAW) screening intervals (abstract).J Urol. 1998; 159: 177Abstract Full Text Full Text PDF Google Scholar demonstrated that after 3 years of follow-up in 8187 men, prostate cancer detection rates were closely linked to the baseline PSA level. For men with baseline PSA levels of 2.0 ng/mL or less, prostate cancer detection rates were 6% or less 3 years later; for men with baseline PSA levels of 2.1 to 4.0 ng/mL, prostate cancer detection rates were 17% 3 years later. Thus, it would appear that the frequency of PSA testing should be linked to the baseline PSA level. Although no upper age limit has been established after which PSA testing is not recommended, there is general agreement that men with less than a 10-year life expectancy are unlikely to gain years of life from early detection because of the long natural history of untreated localized prostate cancer and competing causes of death.30Albertsen P.C Hanley J.A Gleason D.F et al.Competing risk analysis of men aged 55–74 years at diagnosis managed conservatively for clinically localized prostate cancer.JAMA. 1998; 280: 975-980Crossref Scopus (623) Google Scholar In addition, treatment of prostate cancer at older ages is associated with increased morbidity from treatment.31Lu-Yao G.L Albertsen P Warren J et al.Effect of age and surgical approach on complications and short-term mortality after radical prostatectomy—a population-based study.Urology. 1999; 54: 301-307Abstract Full Text Full Text PDF Scopus (81) Google Scholar Thus, even healthy men between 70 and 75 years are reaching an age where further PSA testing may not be beneficial in terms of lives saved, and treatment is associated with an increased chance of quality-of-life reduction. The average life expectancy for United States adults is shown in Table II.32Ventura S.J Peters K.D Martin J.A et al.Births and deaths United States, 1996.Mon Vital Stat Rep. 1997; 46: 1-40PubMed Google Scholar On the basis of these data, PSA testing and treatment of early stage prostate cancer after age 75 is very unlikely to extend life. Nonetheless, more than 20% of the Medicare population is still being tested at age 85.33Health Care Financing Administration, Bureau of Data Management and Strategy, 1998.Google Scholar Given the predictive value of PSA (see section on “PSA as a Predictor of Future Risk”) for the later development of prostate cancer, it would seem reasonable to use PSA measurements to identify older men at low risk of a future prostate cancer diagnosis who may not benefit from intensive screening.TABLE IIAverage life expectancy for U.S. men in 1996legendData from Ventura et al.32Age (yr)Average Number of Years of Life Remaining at Beginning of Age (Life Expectancy)White MenBlack MenAll Races6515.813.915.77012.611.212.5759.89.09.8807.37.07.3legend Data from Ventura et al.32Ventura S.J Peters K.D Martin J.A et al.Births and deaths United States, 1996.Mon Vital Stat Rep. 1997; 46: 1-40PubMed Google Scholar Open table in a new tab Carter et al.34Carter HB, Landis PK, Metter EJ, et al: Prostate specific antigen (PSA) testing of older men. J Natl Cancer Inst (in press).Google Scholar evaluated the association among age, baseline PSA, and prostate cancer detection to test the hypothesis that older men with low PSA levels may require less intensive PSA testing because of a reduced chance of prostate cancer detection. Using the data base from the Baltimore Longitudinal Study of Aging, participants who had serial PSA measurements from age 60, 65, and 70 to their time of diagnosis with prostate cancer (patients) or to age 75 (no cancer) were evaluated. A PSA conversion greater than 4.0 ng/mL was used to define the time of cancer detection among patients with a PSA conversion. The observed and estimated (bootstrap analysis) number of men detected with prostate cancer before age 75 by baseline PSA level and age was calculated. All patients with cancer diagnosed in the 15 years after age 60 had a baseline PSA level greater than 0.5 ng/mL; and 14 of 15 patients who would have been detected by a PSA conversion among the 65-year-old cohort had a baseline PSA level of 1.0 ng/mL or more. Thus, patients with cancer diagnosed by age 75 almost always had PSA levels at baseline a decade earlier that were more than 1.0 ng/mL. In this study group, if PSA testing had been discontinued in men age 65 and older with PSA levels of 0.5 ng/mL or less, 100% (95% confidence interval 78% to 100%) of cancers would still have been detected by age 75. If PSA testing was discontinued in men age 65 years and older with PSA levels of 1.0 ng/mL or less, 94% (95% confidence interval 70% to 100%) of cancers would still be detected by age 75 (Fig. 4). These data suggest that by age 65, men with very low PSA levels (0.5 ng/mL or less or 1.0 ng/mL or less) are not likely to be diagnosed with prostate cancer during the next decade (to age 75), and that prostate cancer detection would be affected little, if at all, if these men had less intensive PSA testing. Since an early diagnosis of prostate cancer is not likely to extend life when made after age 75, the target population for PSA testing is younger than age 75. The use of low PSA levels to identify older men for less intensive screening would reduce the number of men undergoing PSA testing unnecessarily and may not significantly affect cancer detection. Although cost-effective screening intervals need to be identified for older individuals at low risk, widened intervals (2 to 5 years) for older men at lower risk and cessation of PSA testing completely in men without a 10-year life expectancy could result in large health care cost savings without affecting the benefits of PSA testing. Widespread PSA testing for the early detection of prostate cancer in the United States can be considered an uncontrolled experiment.35Stephenson R.A Stanford J.L Population-based prostate cancer trends in the United States patterns of change in the era of prostate-specific antigen.World J Urol. 1997; 5: 331-335Crossref Scopus (106) Google Scholar Population trends since the onset of PSA testing demonstrate some of the characteristics of a successful screening program. If a decrease in prostate cancer mortality is proved to be related to the intensity of screening in the United States and in Europe, data to support PSA testing may be available before the completion of randomized trials that were designed to address this controversy. As was the case with cervical cancer screening, studies will then focus on appropriate guidelines that lead to the cost-effective diagnosis of curable disease. Although efforts to define appropriate guidelines for the early detection of prostate cancer have just begun, it is not too early to begin to address these issues as increasing data become available from serial PSA testing of populations.