Prostate-specific antigen concentration in young men: new estimates and review of the literature

Authors

  • Siobhan Sutcliffe,

    Corresponding author
    1. Division of Public Health Sciences
    2. Alvin J. Siteman Cancer Center, Department of Surgery, Washington University School of Medicine, St Louis, MO
      Siobhan Sutcliffe, Division of Public Health Sciences and the Alvin J. Siteman Cancer Center, Department of Surgery, Washington University School of Medicine, 660 S. Euclid Ave., Box 8100, Rm 5026, St Louis, MO 63110, USA. e-mail: sutcliffes@wudosis.wustl.edu
    Search for more papers by this author
  • Ratna Pakpahan,

    1. Division of Public Health Sciences
    Search for more papers by this author
  • Lori J. Sokoll,

    1. Department of Pathology, Johns Hopkins Medical Institutions, Baltimore, MD
    2. the James Buchanan Brady Urological Institute and the Sidney Kimmel Comprehensive Cancer Center, Johns Hopkins Medical Institutions, Baltimore, MD
    Search for more papers by this author
  • Debra J. Elliott,

    1. Department of Pathology, Johns Hopkins Medical Institutions, Baltimore, MD
    Search for more papers by this author
  • Remington L. Nevin,

    1. Department of Preventive Medicine, Bayne-Jones Army Community Hospital, Folk Port, LA, USA
    Search for more papers by this author
  • Steven B. Cersovsky,

    1. US Army Institute of Public Health, US Army Public Health Command (Provisional), Aberdeen Proving Ground, MD
    Search for more papers by this author
  • Patrick C. Walsh,

    1. the James Buchanan Brady Urological Institute and the Sidney Kimmel Comprehensive Cancer Center, Johns Hopkins Medical Institutions, Baltimore, MD
    Search for more papers by this author
  • Elizabeth A. Platz

    1. Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, and the James Buchanan Brady Urological Institute and Sidney Kimmel Comprehensive Cancer Center, John Hopkins Medical Institutions, Baltimore, MD
    Search for more papers by this author

Siobhan Sutcliffe, Division of Public Health Sciences and the Alvin J. Siteman Cancer Center, Department of Surgery, Washington University School of Medicine, 660 S. Euclid Ave., Box 8100, Rm 5026, St Louis, MO 63110, USA. e-mail: sutcliffes@wudosis.wustl.edu

Abstract

Study Type – Diagnostic (cohort)

Level of Evidence 2b

What's known on the subject? and What does the study add?

Although non-recommended PSA testing has been reported in men younger than 40 years of age, there are few recognized data on PSA in younger American men, particularly younger African-American men, to provide age- and race-specific references.

Using data from an existing large study of young, male members of the US military, aged 28–36 years, the present study provides PSA reference distributions for young Caucasian-American men (median = 0.56, 95th percentile = 1.42, range: <0.01–3.34 ng/mL) and African-American men (median = 0.64, 95th percentile = 1.89, range: 0.12–6.45 ng/mL). Previous estimates from the literature are also summarized.

OBJECTIVE

  • • To provide race-specific prostate-specific antigen (PSA) reference distributions for young men less than 40 years of age who might have undergone non-recommended PSA testing because of their family history of prostate cancer or inadvertently as part of a standard panel of tests.

MATERIALS AND METHODS

  • • We used data from a large existing study of young, male Caucasian- and African-American members of the US military with stored serum in the Department of Defense serum repository.
  • • As part of this previous study, we selected a random sample of 373 Caucasian- and 366 African-American men aged 28–36 years with an archived serum specimen collected for standard military purposes from 2004 to 2006.
  • • We measured serum total PSA concentration in this specimen using the Beckman Coulter Access Hybritech PSA assay.

RESULTS

  • • The PSA level ranged from <0.01 to 3.34 ng/mL among Caucasian-American men, with a median of 0.56 ng/mL and a 95th percentile of 1.42 ng/mL.
  • • The PSA level ranged from 0.12 to 6.45 ng/mL among African-American men, with a median of 0.64 ng/mL and 95th percentile of 1.89 ng/mL.
  • • The PSA level was significantly higher in African- than in Caucasian-American men (P= 0.001).

CONCLUSION

  • • The PSA estimates, together with those summarized from the literature, provide age- and race-specific PSA reference distributions for young men who might have undergone non-recommended PSA testing.
  • • Comparisons by race could also begin to inform the timing of divergence of prostate cancer risk by race.
Abbreviations
DoDSR

Department of Defense serum repository

HIV-1

HIV type 1

INTRODUCTION

There are conflicting guidelines for PSA screening for prostate cancer. While the US Preventive Services Task Force currently recommends against PSA screening [1], other organizations, such as the AUA [2], the National Comprehensive Cancer Network [3], and the European Association of Urology [4], recommend or suggest a risk-based screening strategy for prostate cancer, whereby informed men undergo baseline PSA testing beginning in their 40s as a means of stratifying them into those at higher and lower prostate cancer risk. Anecdotally, PSA testing has also been reported in men even younger than 40 years of age because of their family history of prostate cancer or because their physician indiscriminately ordered a standard panel of laboratory tests. Despite these anecdotal reports, however, there are few recognized data on PSA in younger American men, particularly younger African-American men, to provide age- and race-specific references [5–8]. Therefore, to further inform PSA distributions in young American men, we analysed data from a large existing study of young, male Caucasian- and African-American members of the US military under 40 years of age with stored serum in the Department of Defense serum repository (DoDSR). In contrast to studies of younger men seeking PSA screening, participants in the present study provided blood samples as part of routine military procedures and thus should be unselected for factors, such as a family history of prostate cancer, that might potentially influence their PSA concentration [9]. Finally, in addition to providing estimates of PSA concentration in young Caucasian- and African-American men, our comparison of estimates by race could also begin to inform the timing of divergence of prostate cancer risk by race, as early-adulthood PSA concentrations have been found to predict later incidence of prostate cancer, including aggressive and lethal disease [10–12].

MATERIALS AND METHODS

Since the early 1990s, all members of the US military have undergone routine HIV type 1 (HIV-1) testing as part of the military's HIV-1 surveillance programme [13]. Serum remaining from routine testing is stored in the DoDSR, together with residual serum from indicated HIV-1 testing and pre- and post-deployment sero-surveillance. As part of a separate investigation on PSA in the DoDSR, we selected a random sample of 750 Caucasian- and African-American men who met the following criteria: they were HIV-1-negative; they were <25 years of age as of 1995; they had served on continuous active duty from 1995 to 2006; and they had at least one serum specimen collected between 2004 and 2006. For the present study, we excluded men with a diagnosis of BPH or chronic prostatitis before blood draw (n= 7), as well as those with an acute inflammatory, genitourinary diagnosis in the month before blood draw (n= 3). One man with insufficient serum for PSA testing was also excluded, leaving a final sample size of 739.

The present study was approved by the institutional review boards at the Walter Reed Army Institute of Research and the Johns Hopkins Bloomberg School of Public Health. All data and specimens were anonymized before release from the DoDSR.

Serum specimens collected between 2004 and 2006 were stored at −30°C for 0.7–3.7 years (median = 1.4 years) in the DoDSR and then at −70 to −80°C for 3–15 months until PSA testing. Specimens underwent one freeze–thaw cycle before being shipped from the DoDSR. We performed PSA testing at Johns Hopkins because participants were too young (28–36 years of age) for routine prostate cancer screening and thus did not have PSA concentrations in their medical record. We measured serum total PSA using the Access Hybritech PSA assay (Beckman Coulter, Brea, CA, USA). We tested anonymized specimens in random order and determined assay reproducibility by including 25 blinded quality control pairs from the DoDSR in the testing sequence (coefficient of variation = 12.4% and 6.9% after excluding one discrepant pair).

For the statistical analysis, we calculated separate PSA distributions for Caucasian- and African-American participants by age because these factors are known to influence PSA in older men [14,15]. We compared PSA distributions by race using both a non-parametric Wilcoxon rank-sum test and linear regression with natural log-transformation of PSA. We investigated possible interactions between race and age with PSA by including an interaction term in the linear regression model and evaluating the significance of this term by the likelihood ratio test. All analyses were performed using SAS version 9.2 (SAS, Cary, NC, USA).

RESULTS

In total, 373 Caucasian- and 366 African-American men between the ages of 28 and 36 years were included in the analysis; 87% were enlisted men and 13% were officers. PSA level ranged from <0.01 to 3.34 ng/mL among Caucasian-American men, with a median of 0.56 ng/mL and a 95th percentile of 1.42 ng/mL (Fig. 1). Among African-American men, PSA level ranged from 0.12 to 6.45 ng/mL, with a median of 0.64 ng/mL and a 95th percentile of 1.89 ng/mL. PSA level was significantly higher in African-American than in Caucasian-American men (P= 0.004 by the Wilcoxon rank-sum test; P= 0.001 by linear regression with natural log-transformation of PSA). Among African-American men, PSA increased non-significantly with age (P= 0.052), whereas among Caucasian-American men, there was no association (P= 0.94); however, this suggestion of an interaction was not statistically significant (P interaction = 0.23).

Figure 1.

Distribution of PSA concentration in young adult men in the US military (2004–2006). For each box plot, the lowest line represents the smallest value; the lower end of the box represents the 25th percentile (P25); the line within the box represents the median (P50); the upper end of the box represents the 75th percentile (P75); the upper line at the top of the whisker represents the largest value below the upper fence [the sum of the 75th percentile and 1.5 times the interquartile range (the difference between the 25th and 75th percentiles]); the circles represent values between the upper fence and the far upper fence (the sum of the 75th percentile and three times the interquartile range); and the asterisks represent values greater than the far upper fence.

DISCUSSION

In this large study of young Caucasian- and African-American members of the US military, PSA values ranged from <0.01 to 6.45 ng/mL, with median values of 0.56 ng/mL for Caucasian-American men and 0.64 ng/mL for African-American men. Our values for Caucasian-American men are generally similar to those observed in previous medium- to larger-sized studies of Caucasian-American and Caucasian -European men, including those performed specifically to investigate age-specific PSA distributions [5,6,16–19], and those that presented age-specific distributions in the context of other research questions (Table 1) [20–28]. Our PSA values are higher, however, than those observed for both Caucasian- and African-American men in two more recent, large American studies [7,11], one of which used serum samples from the same broader population as our study (i.e. DoDSR), but from an entirely different sample of men. This independent DoDSR study used samples stored under the same conditions and for a similar length of time as our samples, as well as a PSA assay from the same family of Hybritech assays as our study, although higher PSA values have been observed for the Hybritech Access than for the Tandem assay in a previous multicentre evaluation study [29].

Table 1.  Distribution of serum total PSA concentration in studies of young adult Caucasian and African-American men (≤40 years of age)*
Reference, yearStudy populationAge, yearsSample sizePSA assayMin.PSA concentration, ng/mL
P5P25P50Geometric meanMeanP75P95Max.≥2.5, %
  1. Min., minimum; P5, 5th percentile; P25, 25th percentile; P50, median; P75, 75th percentile; P95, 95th percentile; max., maximum. *Only studies that provided estimates for participants ≤40 years of age were included. Studies with a mean age ≤40 years but without a stated age range or with a maximum age > 40 years were not included unless they provided separate estimates for men ≤40 years of age. Studies of men selected for factors related or possibly related to PSA were also not included (e.g. studies of young men seeking prostate cancer screening or tested for non-study purposes, as these men were likely screened because of a family history of prostate cancer; studies of men with infertility/suspected infertility or spinal cord injuries; and studies composed entirely of obese men). †Nationality assumed based on the affiliation of most study authors. ‡Race assumed based on the geographic location of the study and lack of racial details. §Five men with PSA concentrations more than 3 SD above the mean were excluded from the entire population of men aged 21–76 years. ¶Derived from data presented in the manuscript.

Caucasian (or predominantly Caucasian) men
Stamey et al., 1987 [5]Normal American men†‡§21–3013Pros-Check, Yang labs     0.65    
31–40361.20
Glenski et al., 1992 [6]Healthy American men20–29100Tandem R, Hybritech<0.1  0.6 0.68 1.42.61.0
Wallace et al., 1993 [20]Healthy Scottish men23–4030Tandem R, Hybritech     0.7    
Oremek & Seiffert, 1996 [21]Clinically healthy German outpatients†‡<2537Tandem E, Hybritech       1.4  
25–29321.7
30–39462.2
Heidenreich et al., 1997 [22]Healthy German men25–35100Hybritech assay0.1    0.85  2.2 
Swain, 1997 [50]Healthy American cyclists (participants in a 51–160 km recreational ride)†‡27–399IMx, Abbott0.5  0.70.660.67  0.90.0
Cooper et al., 1998 [23]Healthy American men21–3931Tandem E, Hybritech     0.50–0.83    
Lein et al., 1998 [16]Healthy men from nine European and eight non-European countries20–2941Enzymun-Test, Boehringer Mannheim Diagnostics  0.070.21  0.641.16  
30–39710.390.630.971.78
Chautard et al., 1999 [17]French candidates for a systematic health examination20–29202AxSym, Abbott   0.48 0.55    
30–392180.470.63
Wolff et al., 1999 [18]Rural German men20–2985Immulite, DPC Biermann0.05 0.340.49  0.93 5.2 
30–39860.050.340.681.19.8
Roehrborn et al., 1999 [24]Normal American men†‡<3075Hybritech0.1    0.6  1.7 
30–39920.10.72.2
Preston et al., 2000 [7]Healthy men in the US military20–29253Tandem E, Hybritech  0.270.38 0.470.571.08  
30–392420.280.450.550.681.26
Bhasin et al., 2001 [25]Healthy American men18–3554Immunoradiometric assay     0.5–1.0    
Hartzell et al., 2001 [8]Healthy US military/Department of Defense students30–3921AxSYM, Abbott0.37  0.9 0.89 1.62.0 
Kratz et al., 2003 [51]American amateur marathon runners34–394Elecsys 1010, Roche0.5  0.95 1.4  3.225.0
Linton et al., 2003 [26]Healthy American men<3037Tandem-MP, Hybritech0.21    0.63  0.9 
Pannek et al., 2003 [52]Healthy German men35–4014Tandem E, Hybritech     1.2    
Whittemore et al., 2005 [11]American incidence-density sampled controls in a study of prostate cancer<30127Immunometric kit 0.130.230.32  0.500.84  
30–392070.130.260.370.520.96
Xu et al., 2010 [27]Healthy men in the Swedish military18–21302ProStatus, DELFIA Reagents  0.350.5 0.640.67   
Borgermann et al., 2011 [19]German men recruited through occupational physicians of large companies≤2430Hybritech0.26  0.39 0.53  1.180
25–303840.140.600.631.850
31–355360.170.620.672.561
0.100.580.703.472
36–39828
Korrovits et al., 2011 [28]Estonian mandatory military service candidates and soldiers16–25490DPC Immulite 2000, Siemens0.070.52 4.19 
African-American men
Preston et al., 2000 [7]Healthy men in the US military20–29247Tandem E, Hybritech  0.260.38 0.510.611.35  
30–392470.320.450.570.671.43
Whittemore, 2005 [11]American incidence-density sampled controls in a study of prostate cancer<3072Immunometric kit0.080.220.31 0.520.96
30–39970.120.230.330.551.10
Populations with both Caucasian and African-American men
Baillargeon et al., 2005 [53]Healthy American men (50.1% non-Hispanic white, 36.5% Hispanic white, and 12.4% African-American in the full study population 20 to ≥70 years of age)20–39133IMx, Abbott    0.570.65    

Despite differences in absolute PSA values between the present study and two more recent American studies, there were similar patterns of PSA by race. In the present analysis, we observed higher PSA values among African-American than among Caucasian-American men. These findings are similar to those from the previous independent DoDSR study, which found higher PSA values among African-American than among Caucasian-American men, particularly at the upper end of the PSA distribution [7]. The findings of the present study are also similar to those from a previous nested case–control study of PSA and prostate cancer risk, which found higher 95th percentiles of PSA among young African-American than among Caucasian-American controls, although the statistical significance of this association is unknown [11]. These differences in PSA level between young African-American and Caucasian-American men mirror observed differences in PSA among middle-aged and older African-American and Caucasian-American men [14], as well as differences in prostate cancer incidence and mortality by race [30]. Given that early-adulthood PSA concentrations have been found to predict future prostate cancer incidence [10–12], these findings could suggest that differences in prostate cancer risk between African-American and Caucasian-American men begin to develop before the third or fourth decade of life. By contrast, had we not observed differences in early-life PSA by race, these hypothetical null findings might have suggested that differences in prostate cancer risk by race begin to develop after the fourth decade of life. Finally, although there was no difference in PSA level between African-American and Caucasian-American men aged 30–39 years in a fourth study, these men were selected from voluntary prostate cancer screening programmes [31], and thus which could have been enriched for men with a family history of prostate cancer. Therefore, African-American and Caucasian-American participants might have been more similar with respect to family history of prostate cancer than men sampled for non-screening reasons, such as in our study and the other two studies mentioned [7,11].

Comparing the findings of the present study with those from studies of young Eastern Asian men, the PSA values of the present study were generally lower than those observed in previous studies of Chinese and Korean men (Table 2) [32–40]. This difference has been noted previously [34,35,40,41] and is intriguing because it reflects the opposite relation observed in older men, i.e. lower PSA values in older Asian than Caucasian-American men [34,35,37,40,41], and the opposite relation with prostate cancer risk [42]. However, as no studies have, to our knowledge, collected and tested specimens from Eastern Asian and American men in the same manner and as part of the same study, it is difficult to discern whether differences in PSA reflect biological differences, or simply differences in the PSA assays used, testing conditions or storage conditions. Future studies of Eastern Asian and American men with specimens collected, stored and tested in the same manner would clarify this issue. With respect to PSA concentrations in men of other Asian origins, too few studies of men of South-Central, Western, and other Asian origins have been conducted to adequately compare with our findings.

Table 2.  Distribution of serum total prostate-specific antigen (PSA) concentration in studies of young Asian adult men (≤40 years of age)*
Reference, yearStudy populationAge, yearsSample sizePSA assayMin.PSA concentration, ng/mL
P5P25P50Geometric meanMeanP75P95Max.≥2.5, %
  1. Min., minimum; P5, 5th percentile; P25, 25th percentile; P50, median; P75, 75th percentile; P95, 95th percentile; Max., maximum. *Only studies that provided estimates for participants ≤40 years of age were included. Studies with a mean age ≤ 40 years of age but without a stated age range or with a maximum age > 40 years were not included unless they provided separate estimates for men ≤40 years of age. Studies of men selected for factors related or possibly related to PSA were also not included (e.g. studies of young men seeking prostate cancer screening or tested for non-study purposes, as these men were probably screened because of a family history of prostate cancer; studies of men with infertility/suspected infertility or spinal cord injuries; and studies composed entirely of obese men). †Nationality assumed based on the affiliation of most study authors. ‡Derived from data presented in the manuscript.

Eastern Asian
Lin et al., 1996 [32]Healthy Chinese men21–3070ELISA-PSA2, CIS International     1.04 1.92  
31–402111.031.85
Kao, 1997 [33]Healthy Chinese men<4097PSA-RIACT, CIS Bio International0.310.340.520.71 0.830.941.504.74 
Lee et al., 2000 [34]Healthy Korean men30–391382ELSA-PSA2, CIS Bio International 0.20.60.8  1.11.8  
Ku et al., 2002 [35]Healthy Korean men20–29225Tandem R, Hybritech 0.330.610.90  1.282.25 2.6
30–391 8670.300.600.891.292.354.3
He et al., 2004 [41]Healthy Chinese men23–2977ELSA-PSA, E & E LABS   0.50   1.20  
30–39189 0.55 1.21
Ahn & Ku, 2006 [36]Healthy Korean men20–392 032Tandem R, Hybritech  0.520.85  1.25   
Choi et al., 2007 [37]Healthy Korean men30–3927 553AxSYM and Architect i2000, Abbott, and Elecsys 2010, Roche0.320.550.680.871.021.88
Chang et al., 2008 [38]Healthy Korean men≤30616Tandem R, Hybritech    1.131.12    
31–351 2581.091.11
36–401 7401.141.14
Park et al., 2009 [39]Healthy Korean men30–395 732Not stated    0.93     
Yuan et al., 2011 [40]Healthy Chinese men30–391 135Electrochemiluminescence immunoassay, Roche 0.340.570.80  1.111.89  
South-Central Asian men
Malati & Kumari, 2004 [54]Healthy Indian men20–2940Enzymun-Test, Roche00.6 0.8 0.66 1.071.7 
30–395200.840.90.971.132.2
Western Asian
Kirkali et al., 1995 [55]Normal Turkish men21–2518Tandem R, Hybritech0.09 0.460.700.640.791.13 1.610.0
Kehinde et al., 2005 [56]Healthy indigenous Arab blood donors from Kuwait and Oman15–1952Immulite, DPC     0.32    
20–29680.56
30–39640.49
Saka et al., 2009 [57]Healthy Turkish men17–3524Immulite 2000, Siemens0.09    0.5  1.2 
Populations of men with several different Asian origins
Saw & Aw, 2000 [58]Healthy Singaporean men (78% Chinese, 14% Malay, and 8% Indian origin)30–3979AxSym, Abbott0  0.61   1.36  

As a final consideration, although we excluded men with recent acute inflammatory, genitourinary diagnoses from the analyses, no information was collected in the military medical record on other factors that might influence PSA transiently, such as recent ejaculation or recent bicycle riding, to allow us to test the sensitivity of our findings to these factors. However, neither of these factors has been found to influence PSA consistently or to a great degree, especially in younger men [19,21,43–47]. Therefore, the possible influence of these two factors on both the absolute PSA values and the racial comparison in the present study is debatable. Most previous studies included in Tables 1 and 2 also did not consider these two factors in describing the PSA distributions of their populations. Finally, we did not adjust for BMI as a marker of overweight and obesity, which is also known to influence PSA [48], because members of the US military on active duty are required to maintain their weight below age-specific percentage body fat thresholds and are enrolled in mandatory weight-control programmes if they exceed these limits [49]; therefore, relatively few members of the military are overweight and very few are obese.

In summary, the present study provides new estimates of PSA among Caucasian- American and African-American men of a lower age than those typically studied to date, as well as a summary of estimates from additional young male populations of varying race/ethnicity, to serve as race-specific reference distributions for young men who intentionally or inadvertently undergo PSA testing. PSA comparisons by race could also begin to inform the timing of divergence in prostate cancer risk by race.

ACKNOWLEDGEMENTS

We thank Dr Angelia A. Eick and Zheng Hu at the Armed Forces Health Surveillance Center for help with participant selection, and Dr Catherine G. Sutcliffe for help preparing serum specimens for testing and coordinating PSA testing.

CONFLICT OF INTEREST

None declared. Source of funding: the Patrick C. Walsh Prostate Cancer Research Fund and the Barnes-Jewish Hospital Foundation.

Ancillary